Back to Changelog

    Version 4.0

    Foxit Conversion SDK Changelog v4.0

    Windows/Linux

    October 13, 2026

    New features and enhancements​

    PDF ↔ Markdown​

    • Added PDF to Markdown: foxit::conversion::pdf2md::PDF2Md::ConvertToMarkdown, configurable via PDF2MdSettingData (table format, headers/footers, bold/italic, skip scanned pages, and more). Requires the PDF2MD module license.

    • Added Markdown to PDF: foxit::conversion::md2pdf::Md2PDF::ConvertFromMarkdown, configurable via Md2PDFSettingData (page size and margins, base font size, custom font path, style JSON, font embedding, and more). Requires the MD2PDF module license.

    PDF to Office​

    • Improved dual-layer / scanned PDF to Word performance and image-only page handling

    • Reduced unnecessary PDF to PPT output file-size growth

    • Improved Lao, Myanmar, Khmer, and Thai font support in PDF to Word. Suggested default fonts: Lao DokChampa, Myanmar Myanmar Text, Khmer MoolBoran, Thai Angsana New (DokChampa, MoolBoran, and Angsana New usually require a separate install)

    Bug fixes​

    PDF to Word​

    • Fixed conversion failure returning error code 103 on specific files

    • Fixed errors/crashes on Linux when combining OCR with the PDF2Word engine

    • Fixed possible hangs during PDF to Word

    • Fixed failure handling bold for MICROSS.TTF

    • Fixed loss of bold on subsection titles after conversion

    • Fixed incorrect font substitution for PDF to Word/Excel

    • Fixed leftover newline placeholders and errors at page bottoms on Linux C++

    • Fixed incorrect line breaks

    • Fixed blank pages and abnormal layout in some results

    • Fixed residual newline artifacts affecting downstream translation layout

    • Fixed text misclassified as shapes; TOC/bookmarks not converted to heading structure

    • Fixed incomplete remediation of prior Word conversion regression cases

    • Fixed vector regions incorrectly converted to tables

    • Fixed PDF/A to Word converting text into images

    • Fixed borderless tables becoming text boxes, and slow ML recognition

    • Fixed text conversion mismatches vs the source PDF

    • Improved overall PDF to Word quality on specific samples

    • Fixed letter “I” misrecognized as a Roman-numeral bullet

    PDF to Excel​

    • Fixed crashes on specific files with e_WorkbookSettingsEachPage or e_WorkbookSettingsEachTable

    • Improved poor PDF to Excel quality caused by the underlying engine

    • Fixed cell fill colors exported as images instead of fill

    Word to PDF​

    • Improved overall Office to PDF quality

    • Fixed font weight and horizontal-line position mismatches vs the source

    • Fixed missing nested tables

    • Fixed incorrect table numbering

    • Fixed missing glyphs, extra spaces, and header page-count errors

    • Fixed text color changes, missing content, and incorrect table structure

    • Fixed rewritten heading numbers and inconsistent page counts

    • Fixed layout errors that inflated page counts

    Excel to PDF​

    • Fixed radio buttons not displayed

    • Fixed image stacking/offset covering document content

    • Fixed image stacking offset

    • Fixed image offset stacking and punctuation mojibake

    PowerPoint to PDF​

    • Improved PPTX to PDF output quality