Foxit PDF SDK Changelog

Web

April 21, 2026

Overview

This release delivers 17 enhancements across bookmarks, text extraction, annotations, forms, signatures, 3D and measurement, and host integration. It also includes 60 customer-reported bug fixes. If you upgrade from 11.0, review Bookmarks under Major Changes first.

Major Changes

Bookmarks — async destinations (breaking)

What changed

  • On BookmarkData, page, left, top, zoomMode, and zoomFactor are deprecated as of 11.1.0. The initial bookmark tree no longer embeds full destination payloads for performance on large and linearized / range-loaded PDFs.

What to use instead

  • BookmarkDataService.getBookmarkDestination(bookmarkId) — async destination data (coordinates, zoom, and related fields) for thumbnails, overlays, or custom scroll logic.
  • BookmarkDataService.performAction(bookmarkId) — executes the bookmark’s PDF action chain (GoTo, URI, JavaScript, and so on), matching the built-in bookmark panel. You do not need to call getBookmarkDestination before performAction for standard navigation.
  • Prefer page object number when linking bookmarks to document pages instead of relying only on a volatile pageIndex.

New Features and Enhancements

Page text and Advanced Edit

  • PDFPage: getText, getCharCount, getCharInfoByIndex, and getCharRange support character-index workflows (for example proofreading, anchors for comments, and automation alongside search APIs).
  • PDFPage.addRichText: programmatic rich text on the page (requires AdvEdit license).

Annotations and forms

  • PDFViewer.setMultipleSelectionEnabled / isMultipleSelectionEnabled — enable or disable marquee (rectangular drag) selection, Ctrl/Shift-click multi-select, and Ctrl+A / Cmd+A select-all for annotations and form fields.
  • PDFPage.setAnnotGroup and MarkupAnnot.ungroup — group or ungroup markups without creating new PDF objects; stable objNumber / name. Listen for AnnotUpdatedType.groupUpdated in annotationUpdated for group-related UI state.
  • enableRealtimeAnnotShapeUpdate (on/off) — optional realtime refresh of markup shape while dragging or resizing. Set it via PDFViewer constructor options and/or pdfViewer.config.enableRealtimeAnnotShapeUpdate at runtime.
  • AnnotComponentConfig.enableDiagonally — diagonal resize handles for Callout when used with annot component configuration callbacks.
  • FreeText — editing behavior closer to Foxit PDF Editor (desktop) for mixed scripts and layout.
  • Form push button — Select Icon modal: browse, preview, and clear the button image (JPEG, PNG, GIF, BMP).

Signatures and stamps

  • Quick Sign (under Protect) — on-page PSI strokes with pressure-style rendering, alongside classic PDF Sign dialog drawing.
  • Signature UI — clear signature from context menu; Sign & Certify with place signature; PDFDoc.sign extended with isKeepRatio and custom bitmap overlay for appearance.
  • Custom dynamic stamps — multi-label configuration in the UI and via CustomDynamicStamp / setDynamicStamp (field as object or array).

3D and measurement

  • 3D PDF — sidebar model tree (parts visibility control) and view list with preset cameras for design-review style navigation.
  • CreateMeasurementService (from pdfViewer.getCreateMeasurementService()) — programmatic control of measurement creation (for example markup-on-create, label, results panel visibility, snap modes, scale). Replace ad hoc DOM coupling.

Viewer integration

  • PDFViewer constructor — customs.containerRoot (HTMLElement) to scope styles when embedding without an iframe (works with Shadow DOM patterns).
  • Stable component-name attributes – on annotation menus and related components for predictable getComponentByName customization.

Performance Optimizations

  • Bookmarks — async getBookmarkDestination and a slimmer initial tree reduce up-front work and improve bookmark panel responsiveness on large linearized documents and openPDFByHttpRangeRequest loads.
  • Rendering – by significantly optimizing the rendering pipeline, the zoom interaction speeds have been boosted by up to 39%， and annotation & form page navigation are up to 34% faster.

Bug Fixes

Rendering

  • Fixed page.render bitmap export using a black background in SDK 11 instead of the 10.x white background.
  • [Performance]Fixed mid-batch page.render image conversion failing with getPageBitmap engine errors.
  • [Performance]Fixed very large drawings causing excessive memory or crashes on mobile and heavy memory or render stalls on desktop.
  • Fixed brief desync between annotations and selection chrome during zoom (“drift / judder”).
  • Fixed occasional whole-page shift after pinch-zoom release on phones.
  • [Performance]Fixed Safari scroll jank, sidebar affecting first paint and navigation, and non-passive wheel listener performance warnings.
  • [Performance]Reduced stamp flicker while zooming the PDF.
  • [Performance]Improved render throughput when continuously scrolling large documents after rotating the view.

Document and page

  • Fixed blank UI or “file damaged” when using loadPDFDocByHttpRangeRequest before a fragmented openPDFByHttpRangeRequest open.
  • Fixed collaboration where the first user closing the co-editing page left other users in the same session URL with missing or unsynced annotations.
  • [Performance]Fixed getEmbeddedFile taking too long on very large embedded attachments and the browser crashing on extremely large files.
  • Fixed openPDFByHttpRangeRequest on specific PDFs showing a blank virtual page 2 and first-page interaction throwing Unknown engine error.

Text and content

  • Fixed compareDocuments under Vietnamese (vi-VN) dropping summary or result title text in the output comparison PDF.
  • Fixed PDFDoc.searchText regex search (Search & Redact) failing on very high page-count PDFs.
  • Fixed Advanced Edit text mode with Japanese IME: full-width brackets left the selection stuck and impossible to clear.
  • Fixed redaction selection offset when applying redactions from search hits after alternating search and manual redact (text index not updated after text removal).

Forms and XFA

  • Fixed “Create multiple copies” vertical and horizontal counts inverted versus 10.x and desktop Editor.
  • Fixed dynamic XFA form export to XML then re-import preventing field values from writing back.
  • Fixed hand tool clicks on form fields still firing activeAnnotation.
  • Fixed GDOR tax sample issues: memory error when closing the help dialog, Tab key paging incorrectly, and date widget panel not auto-collapsing.

Annotations — interaction, display, and permissions

  • Fixed Callout on specific documents: could not place on the left of the page; leader or text box wrong or collapsed when creating in succession.
  • Fixed line spacing and last-line mismatch for annotations before versus after blur when using custom mapped fonts.
  • Fixed FreeText box width shrinking abnormally after continuous input at high zoom when the page extends beyond the viewport.
  • Fixed mobile workflow where, after opening a single-page document then switching to a multi-page PDF, vertical swipe paging no longer worked.
  • Fixed Typewriter still editing while wheeling across pages so some pages’ annotations were missing from the annotation list.
  • Fixed setDefaultAnnotConfig not applying default colors or properties for Note and similar annotations.
  • Fixed Callout default rectangle too large and leader start not matching the pointer on some draw paths.
  • Fixed occasional runtime error when hovering over existing annotations (reload and developer-tools-related repro paths).
  • Fixed delete-only permission still allowing move via context-menu align, distribute, or uniform size.
  • Fixed fewer annotations loaded than Adobe on specific documents, with getAnnots engine errors.
  • Fixed measurement annotations incorrectly allowing geometry (rect) edits by default.
  • Fixed Stamp default properties: create on a rotated page then drag to a normal page caused squashed or wrong scale.
  • Fixed area highlight not stacking above TextBox or Callout, inconsistent with desktop Editor z-order.
  • Fixed ruler UI flicker when setting measurement scale units programmatically.
  • Fixed annotComponent.active() or pdfui.activateElement not firing activeAnnotation.
  • Fixed first programmatic Stamp add where rectangle versus stamp aspect or position was wrong, and wrong behavior after switching documents.
  • Fixed Cloud annotation arcs pointing outward on some draw directions instead of consistently inward.
  • Fixed regression where custom stamp dropdown title or tooltip showed category name instead of stamp name.
  • Fixed error switching from hand tool to a newly created custom dynamic stamp tool.
  • Fixed wrong annotation selected or none selected when clicking quickly and moving the pointer away.
  • Fixed Chromium 86–class browsers (for example 360 Enterprise): pencil and typewriter issues, console errors, and page artifacts after page flip.

Annotations — data exchange

  • Fixed highlight exported to JSON then re-imported snapping to text instead of staying a free region highlight.
  • Fixed import hang or blank page on specific documents.
  • Fixed Typewriter on rotated pages where JSON export and import layout differed from pre-export.
  • Fixed caret text-replacement annotations incomplete after importing PDFBox-exported FDF.
  • Fixed non-breaking space (U+00A0) rendering as Euro symbol after JSON export and import.
  • Fixed specific documents erroring when exporting annotations to JSON via API or UI.
  • Fixed XFDF export using page-rotation attribute names inconsistent with older builds and common interop (rotation / rotate).
  • Fixed JSON import failing when content ended with a stray backslash or other bad escaping.

Annotations — grouping

  • Fixed stretching one grouped member shifting sibling positions incorrectly.
  • Fixed setBorderWidth on a group affecting only the active member while other border APIs behaved on the whole group.
  • Fixed unstable or duplicate annotationUpdated firings for move or line width on grouped or multi-selected annotations.
  • Fixed setBorderInfo on a group affecting one member while setBorderStyle and setBorderWidth affected the whole group.
  • Fixed occasional canvas not refreshing after grouped annotation move or property change until the page was scrolled.

Security and signatures

  • Fixed empty document ballooning in size after adding then deleting a Stamp.
  • Fixed viewport scrolling below the signature field after sign when the field sits in the left half of the page.
  • Fixed custom dynamic stamp list showing seams after refresh versus seamless first paint.
  • Fixed Note pop-up icon visibility changing after open-and-download-only via SDK versus desktop Editor preferences.

Other

  • Fixed Next.js 12 integration with the official sample not loading the SDK (aligned with Next 14 sample behavior).
  • Fixed pdfViewer.offShortcutKey('ctrl+v') not disabling paste.

Windows/Mac/Linux

April 3, 2026

Breaking Changes

The following changes require code updates when upgrading from v11.0. Existing projects that use these APIs will not compile or may behave differently without modification.

ActionCallback — Parameter change and new pure virtual methods

The ActionCallback class has 1 parameter change and 7 new pure virtual methods. Any subclass must be updated.

Parameter change — GetPageWindowRect now requires document and page_index parameters:

// v11.0
virtual RectF GetPageWindowRect() = 0;

// v11.1
virtual RectF GetPageWindowRect(const foxit::pdf::PDFDoc& document, int page_index) = 0;

New pure virtual methods — implement with empty bodies if the functionality is not needed:

virtual void NotifyBeginDoJob(const pdf::PDFDoc& document,
    JavascriptModifyItemInfo::JavascriptEventType event_type) = 0;
virtual void NotifyAfterDataChange(const pdf::PDFDoc& document,
    JavascriptModifyItemInfo modify_item_info) = 0;
virtual void NotifyEndDoJob(const pdf::PDFDoc& document,
    JavascriptModifyItemInfo::JavascriptEventType event_type) = 0;
virtual bool InitModifyItem(const pdf::PDFDoc& document,
    ModifyItemType item_type, int page_index, const WString& dict_name) = 0;
virtual void ResetModifyItem(const pdf::PDFDoc& document) = 0;
virtual int GetVisiblePageCount(const pdf::PDFDoc& document) = 0;
virtual int GetVisiblePage(const pdf::PDFDoc& document, int index) = 0;

DocProviderCallback — New pure virtual method (XFA)

A new pure virtual method is added. Subclasses of DocProviderCallback must implement it:

virtual void NotifyWidgetChangeInfo(const XFADoc& doc, XFAWidgetModifyInfo change_info) = 0;

This only affects projects that use the XFA module. Implement with an empty body if not needed.

IconProviderCallback::GetIcon — Parameter change (silent)

The GetIcon method has changed. This is not a pure virtual, so compilation will succeed, but existing overrides will silently stop being called:

// v11.0
virtual PDFPage GetIcon(Annot::Type annot_type, const char* icon_name, ARGB color);

// v11.1
virtual PDFPage GetIcon(Annot::Type annot_type, const char* icon_name, ARGB color,
                        foxit::pdf::objects::PDFDictionary* annot_dict);

Action required: Update the override to include the new annot_dict parameter. Enable -Woverloaded-virtual (GCC/Clang) or /W4 (MSVC) to detect this at compile time.

OCRCallback — New pure virtual method

A new pure virtual method IsImageIgnored is added to the existing OCRCallback class:

// v11.0
class OCRCallback {
  virtual bool NeedToCancelNow(const wchar_t* info) = 0;
};

// v11.1
class OCRCallback {
  virtual bool NeedToCancelNow(const wchar_t* info) = 0;
  virtual bool IsImageIgnored(foxit::pdf::graphics::ImageObject* image_object) = 0; // NEW
};

Action required: All subclasses of OCRCallback must implement IsImageIgnored. Return false to preserve previous behavior (no images ignored).

OCRConfig constructor and Set() — Parameter change

The parameterized constructor and Set() method gained 3 new required parameters:

// v11.0 (5 parameters)
OCRConfig(bool is_detect_pictures, bool is_remove_noise, bool is_correct_skew,
          bool is_enable_text_extraction_mode, bool is_sequentially_process);

// v11.1 (8 parameters)
OCRConfig(bool is_detect_pictures, bool is_remove_noise, bool is_correct_skew,
          bool is_enable_text_extraction_mode, bool is_sequentially_process,
          bool is_auto_overwrite_resolution, int resolution_to_overwrite, int confidence);

Action required: Add the 3 new parameters. Use true, 300, 0 for default-equivalent behavior. Alternatively, use the default constructor which initializes all fields with defaults.

OCR::OCRPDFPage / OCRPDFDocument / OCRConvertTo / OCRPDFDocuments — New parameter

All OCR processing methods add an OCRProgressCallback* parameter:

// v11.0
void OCRPDFPage(PDFPage pdf_page, bool is_editable);
void OCRPDFPage(PDFPage pdf_page, bool is_editable, const OCRConfig& config);
// ... same pattern for OCRPDFDocument, OCRConvertTo, OCRPDFDocuments

// v11.1
void OCRPDFPage(PDFPage pdf_page, bool is_editable, OCRProgressCallback* callback = NULL);
void OCRPDFPage(PDFPage pdf_page, bool is_editable, const OCRConfig& config, OCRProgressCallback* callback = NULL);
  • C++: Has default value NULL — existing calls compile without changes.
  • C# / Python / Java / Node.js / Go: Bindings do not inherit the C++ default value. Existing calls must add the new parameter explicitly (pass null / None / nil to preserve previous behavior).

PDF to Office — API parameter changes

The following conversion APIs have breaking changes in this release. For detailed migration guidance with code examples, see the Breaking Changes section in the Conversion SDK v3.1.0 Release Note (included in the Conversion SDK package).

  • PDF2WordSettingData — Add required parameter max_blank_paragraphs_per_page_bottom; default value of enable_generate_headers_and_footers changed from false to true
  • PDF2PowerPointSettingData — Add required parameter enable_adapt_to_largest_page
  • PDF2ExcelSettingData — Add required parameters enable_aggressive_table_repair and include_watermarks
  • PDF2OfficeSettingData — Add required parameter enable_matching_system_fonts
  • PDF-to-Word now preserves internal document navigation links by default

New Features and Enhancements

Platform

  • Add Node.js v21 and v22 support (extends previous v18-20 range)
  • Add Node.js macOS platform support

Conversion

  • Add WPS-powered Office-to-PDF conversion on Windows via Convert::FromWord/Excel/PowerPoint(..., e_Office2PdfEngineWps)
  • Add system font precise matching toggle for PDF-to-Word: PDF2OfficeSettingData::enable_matching_system_fonts
  • Update PDF-to-Word to preserve internal document navigation links by default
  • Add page size adaptation for PDF-to-PPT: PDF2PowerPointSettingData::enable_adapt_to_largest_page
  • Add PDF-to-Excel table repair and watermark output controls: PDF2ExcelSettingData::enable_aggressive_table_repair / include_watermarks
  • Add font embedding toggle for Linux Office2PDF: Office2PDFSettingData::is_embed_font
  • Add version query APIs: Office2PDF::GetVersion() / PDF2Office::GetVersion()
  • Update Office2PDF::ConvertFromWord to support DOC (Word 97-2003) format

OCR

  • Add progress callback via OCRProgressCallback class with ProgressNotify(int current_rate)
  • Add Arabic language support: OCREngine::SetLanguages("Arabic")
  • Add configuration parameters: OCRConfig::confidence (confidence threshold) and OCRConfig::resolution_to_overwrite (resolution override)
  • Add standalone command-line tool for multi-process parallel OCR: ocr_win64.exe / ocr_linux64
  • Add image filtering callback: OCRCallback::IsImageIgnored(ImageObject*)

Rendering

  • Add overprint rendering support: Renderer::SetOverprint(bool is_to_enable_overprint)

Security and Signature

  • Add pre-sign self-modification tracking callbacks in ActionCallback: InitModifyItem(), ResetModifyItem(), NotifyBeginDoJob(), NotifyAfterDataChange(), NotifyEndDoJob(); add DocProviderCallback::NotifyWidgetChangeInfo() for XFA widget change tracking
  • Add CertChainResolverCallback and TrustedCertStoreCallback for cross-CA LTV enablement in LTVVerifier
  • Add Redaction::EnableFileStream() to reduce memory peaks in high-redaction-volume scenarios via file stream

Forms and Annotation

  • Add 29 XFA event types to XFADoc::EventType for granular event handling (Click, Change, Enter, Exit, PreSign, PostSign, PreSave, PostSave, etc.)
  • Update IconProviderCallback::GetIcon with annot_dict parameter for annotation dictionary access

Document and Page

  • Update ActionCallback::GetPageWindowRect with document and page_index parameters for multi-page JS layer (OCG) control
  • Add ActionCallback::GetVisiblePageCount() and ActionCallback::GetVisiblePage() for querying visible pages in multi-page view
  • Add methods to remove PDF logical structure tags: PDFDoc::RemoveStructTree(), PDFStructTree::RemoveChild(), StructElement::RemoveChild()
  • Add Font::IsCharSupported(uint32 unicode, const PDFDoc& document) for character support detection
  • Add Library::AddExternalFontPath() and Library::MatchExternalFontsOnly() for external font path management

3D

  • Add PDF3DContext::Add3DAnnot() to insert 3D annotation on a specified page (file path and ReaderCallback modes)
  • Add 3D preset view and model tree interaction: PDF3DAnnotInstance::ApplyPresetView(), GetPresetViewList(), ModelNode class with visibility control

Optimization

  • Add transparency optimization with configurable resolution modes (Low/Medium/High): OptimizerSettings::SetTransparencyMode()

Bug Fixes

Conversion

  • Fix PDF-to-Word crop marks at page corners causing translation software failures
  • Improve PDF-to-Word cross-application rendering consistency between MS Office and WPS Office
  • Fix PDF-to-Excel cell text placed inside shapes instead of cells
  • Fix PDF-to-Excel table borders rendered as bitmaps overlaid on the table
  • Fix PDF-to-Excel excessive conversion time with cells appearing as images
  • Fix PDF-to-PPT page dimension changes with content shrunk to upper-left
  • Fix Office2PDF::ConvertFromWord crash on specific DOCX files
  • Fix incorrect character spacing in Word-to-PDF causing wrong line breaks
  • Fix bold text appearing excessively bold in Word-to-PDF output
  • Fix missing text (last sentence lost) in Word-to-PDF output
  • Fix shape rendering deviations and incorrect text line breaks in Word-to-PDF

OCR

  • [Linux] Fix OCRConvertTo failure (ERR_FREN_NO_PAGES) due to missing Chinese fonts
  • Fix OCRConvertTo error (ERR_IMAGE_LIBJPEG_LIBRARY_RAISED_ERROR) on specific files
  • Fix OCRConvertTo only outputting first page of a 44-page document

Rendering

  • Fix incorrect CMYK-to-ARGB color conversion in Bitmap DIBFormat
  • [Performance] Fix progressive rendering slowdown (17s->24s->30s->40s) with multiple open documents
  • Fix OutputPreview::SetSimulationProfile() not reflecting different ICC profiles
  • Fix missing text in OutputPreview rendering of specific PDFs
  • Fix rendering inconsistencies with Adobe on specific PDF pages
  • Fix crash when saving Bitmap created with e_DIBRgb via Image::SaveAs
  • Fix abnormal rendering when adding PathObject to specific document pages
  • Fix memory leak in Bitmap constructed from buffer where delete() did not free memory

Printing

  • Fix color discrepancies and uneven dot patterns in PrintManager output
  • Fix intermittent crash in multi-threaded PrintManager usage
  • Fix FXPM_AddPDFFromFileToJob and FXPM_SetJobDocumentName not supporting CJK paths
  • Fix FXPM_SetJobDuplex(1) not enabling duplex printing
  • Fix mixed-orientation page content not rotating with page direction
  • Fix SetRotation() causing landscape PDFs to print with portrait text orientation
  • Fix PrintJob only printing first file when multiple files added via AddPDFFromFile

Forms

  • Fix XFA ExportData/ImportData cycle corrupting table headers and digital signatures
  • Fix XFA TextField Widget proliferation causing progressive load time increase and crash
  • Fix C# ViewDemo rendering blank pages for specific XFA documents
  • Fix JavaScript layer (OCG) visibility control having no effect
  • [Regression] Fix ActionCallback trigger regression in v11 where some callbacks did not fire after field modification
  • Fix “Inherit Zoom” bookmark destination only working for the first bookmark

Document and Page

  • Fix crash when calling SaveAs with e_SaveFlagLinearized after StartEmbedAllFonts
  • Fix fxhtml2pdf zombie process after Html2PDF timeout
  • [Performance] Fix StartSplitByFileSize taking over 1 hour for 25,000-page PDFs
  • Fix excessive Redaction memory peaks in high-redaction-volume continuous-processing scenarios
  • [Windows] Fix ComplianceEngine::SetTempFolderPath() writing to executable directory instead of system temp
  • Fix C# TextPage.GetTextInRect returning question marks for certain characters
  • Fix GetEditingTextCaretPosition returning incorrect positions causing misplaced IME candidate window

Other

  • Fix visual artifacts from parallel multi-process Optimizer image compression
  • Fix specific 3D PDF showing blank after clicking 3D annotation area
  • Improve AutoTagging recognition of Figure-type images vs. Adobe Auto-Tag
  • Fix incorrect content matching in Comparison for OCR documents

PDF Conversion SDK 3.1.0

April 3, 2026

Breaking Changes

PDF2WordSettingData — Parameter change and default value change

Constructor and Set() method add a required max_blank_paragraphs_per_page_bottom parameter:

// v3.0 (4 parameters)
PDF2WordSettingData(bool enable_retain_page_layout,
    bool enable_generate_headers_and_footers,
    bool enable_generate_footnotes_and_endnotes,
    bool enable_generate_page_rendered_break);

// v3.1 (5 parameters)
PDF2WordSettingData(bool enable_retain_page_layout,
    bool enable_generate_headers_and_footers,
    bool enable_generate_footnotes_and_endnotes,
    bool enable_generate_page_rendered_break,
    int max_blank_paragraphs_per_page_bottom);  // new

Action required: Add the new parameter when using the parameterized constructor or Set(). Pass -1 for unlimited (matching previous behavior). Alternatively, use the default constructor with field assignment.

Default value change — enable_generate_headers_and_footers changed from false to true. Code using the default constructor will now generate headers and footers. Set enable_generate_headers_and_footers = false explicitly to preserve previous behavior.

PDF2PowerPointSettingData — Parameter change

Constructor and Set() method add a required enable_adapt_to_largest_page parameter:

// v3.0 (1 parameter)
PDF2PowerPointSettingData(bool enable_aggressively_split_sections);

// v3.1 (2 parameters)
PDF2PowerPointSettingData(bool enable_aggressively_split_sections,
    bool enable_adapt_to_largest_page);  // new

Action required: Add the new parameter. Pass false to preserve previous behavior.

PDF2ExcelSettingData — Parameter change

Constructor and Set() method add 2 required parameters:

// v3.0 (3 parameters)
PDF2ExcelSettingData(String decimal_symbol, String thousands_separator,
    WorkbookSettings workbook_settings);

// v3.1 (5 parameters)
PDF2ExcelSettingData(String decimal_symbol, String thousands_separator,
    WorkbookSettings workbook_settings,
    bool enable_aggressive_table_repair,  // new
    bool include_watermarks);             // new

Action required: Add 2 new parameters. Pass true, false to preserve previous behavior.

PDF2OfficeSettingData — Parameter change

Constructor and Set() method add a required enable_matching_system_fonts parameter:

// v3.0 (9 parameters)
PDF2OfficeSettingData(const wchar_t* metrics_data_folder_path,
    bool enable_ml_recognition, const Range& page_range,
    bool include_pdf_comments,
    const PDF2WordSettingData& word_setting_data,
    const PDF2PowerPointSettingData& power_point_setting_data,
    const PDF2ExcelSettingData& excel_setting_data,
    bool enable_trailing_space, bool include_images);

// v3.1 (10 parameters)
PDF2OfficeSettingData(/* same 9 parameters */,
    bool enable_matching_system_fonts);  // new

Action required: Add the new parameter. Pass false to preserve previous behavior.

PDF-to-Word default behavior change

PDF-to-Word conversion now preserves document internal navigation links by default. If downstream workflows depend on link-free Word output, verify conversion results after upgrading.

New Features and Enhancements

PDF to Office

  • Add enable_matching_system_fonts parameter to PDF2OfficeSettingData for system font matching in PDF-to-Word conversion
  • Add enable_adapt_to_largest_page parameter to PDF2PowerPointSettingData to unify slide size by scaling smaller pages to the largest page dimension
  • Add enable_aggressive_table_repair and include_watermarks parameters to PDF2ExcelSettingData to strengthen table structure repair and control whether watermark content is retained in Excel output
  • Improve PDF-to-Word conversion to preserve document internal navigation links (TOC, cross-references) by default
  • Improve PDF-to-Word conversion performance and reduce memory usage for complex documents
  • Improve PDF-to-Word heading recognition accuracy for better Word heading style mapping
  • Improve PDF-to-Word paragraph recognition, page break handling, and cross-page table continuity

Office to PDF

  • Add Linux platform support for Office2PDFSettingData::is_embed_font font embedding
  • Add DOC (Word 97-2003) format support to Office2PDF::ConvertFromWord
  • Add e_ErrNotMatchFont error code for font-missing scenarios in Office-to-PDF conversion

Platform

  • Add Linux ARM64 .NET Core platform support

Bug Fixes

PDF to Word

  • Fix crash when converting specific PDF files to Word on Linux .NET platform
  • Fix conversion failure producing no output file for specific documents
  • Fix exception (0x77574080 ntdll.dll) in MFC-based projects during PDF-to-Word conversion in Debug x86 mode
  • [Linux] Fix incorrect font substitution where Song typeface renders as Hei in converted Word output
  • Improve font selection to avoid Segoe UI Light for better cross-application compatibility with WPS Office
  • Fix font style loss (e.g., bold) in converted Word documents
  • Fix incorrect font size in conversion output causing layout disruption
  • Fix extra spaces between characters in paragraph text of converted Word output
  • Fix title character overlap in converted Word documents
  • Improve layout stability for translation round-trip workflows where longer translated text caused subsequent page disruption
  • Fix crop marks at page corners interfering with third-party translation software
  • Fix table of contents misalignment in converted Word documents
  • Fix layout misalignment in converted Word output for specific files
  • Fix PDF headers and footers incorrectly converted as body text instead of Word header/footer sections
  • Fix extra line breaks at end of text lines causing layout disruption during translation write-back
  • Fix Arabic content loss during PDF-to-Word conversion
  • Improve conversion quality for large documents: reduce layout issues, output file size bloat, and excess page count
  • Improve cross-application rendering consistency between Microsoft Office and WPS Office
  • Fix faded text, abnormal table rendering, content loss, and text overlap for specific files
  • Fix table of contents loss or corruption with headings not recognized in converted output
  • Fix enable_generate_headers_and_footers = true not taking effect in PDF2WordSettingData

PDF to Excel

  • Fix cell text placed inside drawing shapes instead of editable cells
  • Fix excessive conversion time and incorrect table structure with cells appearing as images
  • Fix content loss when converting built-in test documents

Word to PDF

  • Fix crash in Office2PDF::ConvertFromWord when converting specific DOCX files
  • Fix hang with memory spike to 4 GB when converting specific Word files
  • Fix table content loss in specific regions during conversion
  • Improve character spacing and paragraph layout fidelity on Windows
  • Fix is_embed_font = true not taking effect on Windows in Office2PDFSettingData
  • [Linux] Fix blank PDF output on Red Hat systems missing Calibri font
  • Fix incorrect character spacing causing line break differences from original Word document
  • Fix bold text appearing excessively thick in converted PDF output
  • Fix missing last sentence in converted PDF for specific documents
  • Fix graphical shape rendering deviations and incorrect text line breaks

Updated on April 21, 2026