Apryse Server SDK release notes
October 1, 2026
v12.02
For this release, the Server SDK expands what developers can produce from code. Markdown output extends the Document Structure API so content extracted from PDFs lands directly in the format AI pipelines consume, rather than being converted downstream. The Doc Generation APIs add a higher-level layer for building documents programmatically. Both run server-side in the customer’s own environment and feed the document pipeline customers already build on.
Smart Data Extraction (SDE) Markdown Output
The SDE Doc Structure API gains Markdown as an output option alongside the existing structured outputs. The extracted Markdown is built on Apryse’s leading document structure intelligence, preserving meaningful hierarchy and context by preserving elements like headings and lists, rather than flattening content into plain text that can make context disappear. This gives AI and RAG pipelines higher-quality, better-structured data for ingestion, retrieval, and more accurate responses.
The new API method is DataExtractionModule.ExtractToMarkdown().
Preserved structural elements include:
Paragraphs in reading order.
Heading levels 1-6.
Nested bulleted and numbered lists, including decimals.
Tables with headers, merged cells, and content.
Nested tables with inline HTML enabled.
Images with alternative text and hyperlinks.
External and internal links.
Referenced bookmarks as HTML anchors.
Text boxes as block quotes, when appropriate.
Bold, italic, strikethrough, superscript, and subscript.
Industries benefit from using this type of extracted Markdown. Use cases include:
Financial services: Indexing research reports, statements, policies, and regulatory filings in a vector database so analysts and service agents can retrieve precise answers with source context.
Insurance: Converting policy documents, claims files, adjuster reports, and underwriting manuals into structured Markdown for claims copilots and coverage-question assistants.
Legal: Ingesting contracts, case files, discovery documents, and precedent libraries into RAG systems for clause search, due diligence, and matter-specific research.
Healthcare and life sciences: Preparing clinical protocols, medical literature, product documentation, and regulatory submissions for controlled internal knowledge assistants.
Government and public sector: Transforming legislation, policy manuals, forms, and public records into searchable knowledge bases while processing sensitive content within the agency's environment.
See the following documentation for developing this feature:
Document Creation API
This release improves Apryse’s document generation capabilities by expanding the high-level Document Creation API and creating a stronger foundation for end-to-end PDF creation workflows. Today, Apryse supports basic document creation capabilities such as paragraphs, text, tables, lists, page size, and margins, but customers also need a more complete API surface to create professional PDFs from scratch without relying on low-level PDF operations.
An invoice, for example, may need a logo, a line-item table, a footer, and page numbers before it is recognizably an invoice. That set is now available without leaving the high-level API. We've added:
Hyperlinks
Page numbers
Headers and footers
Images
Shapes
Floating elements
Charts
PDF metadata
Improved positioning and layout controls
The Document Creation API, FlowDocument, has new Section, Shape, PageNumber, Chart, and Float classes.
The low-level PDF APIs are untouched. They remain available and unchanged. This is a higher-level layer over the existing layout engine for common document workflows, not a replacement. If you have precise, unusual output requirements, that path is still there, and code built on it keeps working.
Use cases include:
Invoice generation at volume: company branding, line-item tables, page numbers, payment information, metadata, and configured page size and margins.
Analytics and business reports: structured content, charts or chart-like visual elements, branded headers, landscape pages, and custom page numbering.
Contracts and proposals: logos, hyperlinks to terms or support, disclaimers in footers, shapes used as signature placeholders, metadata, and multi-page layout.
Statements, dashboards, labels, and other dynamically generated documents produced from operational systems, CRMs, ERPs, billing systems, and reporting tools.
OEM teams embedding the SDK in their own product and needing programmatic PDF generation as part of it.
See the following documentation for developing this feature:
See our latest Changelog for more Server release information, including new APIs and bug fixes.
See our downloads page to access the most recent versions of our products so you have all of the latest features and functionality available.
July 8, 2026
v12.0
The Server SDK introduces new OCR and XLIFF Reflow updates, enhancing what customers can do with documents entirely inside their own environment. The new OCR engine raises text-extraction accuracy on real-world documents and widens language coverage, turning scanned and image-based files into searchable, structured data. XLIFF Reflow adds right-to-left (RTL) language support, so translation workflows can render Arabic and Hebrew layouts automatically rather than through manual rework. Across both, the work runs server-side on the customer’s own infrastructure and feeds directly into the broader document pipeline they already build on.
OCR
We’re excited to announce the new AI-powered OCR engine for the Apryse Server SDK. It significantly improves accuracy on real-world documents and performs better on low-quality scans and complex layouts. Language coverage expands to include Chinese, Japanese, and Korean, Portuguese, and additional Latin-alphabet languages. Extracted text is available as structured JSON or XML for downstream search, indexing, and AI pipelines – all processed on-premise or server-side, cross-platform, so documents never leave the customer's environment.
Use cases include:
Invoices, receipts, and other financial documents.
Forms and applications.
Scanned records in healthcare and legal workflows.
Older archived PDFs that need to be made searchable.
Feeding extracted, structured data into search and indexing, LLM or RAG pipelines, and internal databases.
The new OCR is the default OCR module, while the pre-existing V11 OCR is now an alternative module. When you download the OCR module you’ll download either the default OCR module or alternative OCR module, depending on which you’d like to use.
See the following documentation for developing this feature:
XLIFF - RTL Language Support
XLIFF Reflow adds RTL language support, letting customers apply RTL translations, Arabic and Hebrew, to left-to-right (LTR) source documents. Page layouts mirror automatically, reading order is preserved across complex designs, and mixed-direction text: numbers, dates, and embedded Latin, follows correct bidirectional rules.
RTL languages don’t only translate differently; they flow differently. Handled manually, this puts teams into cycles of rework, layout fixes, and extended QA that delay launches and raise costs. With our new XLIFF Reflow RTL language support, users see formatting fidelity from extraction to reinsertion; end-to-end XLIFF automation with the same workflow as before, now RTL-aware.
Use cases include:
Translating an LTR source document into an RTL language.
Localizing regulatory materials, customer invoices, or policy statements into RTL languages.
Scaling multilingual PDF production into RTL markets without branching document templates or manual layout correction.
See the following documentation for developing this feature:
See our latest Changelog for more Server release information, including new APIs and bug fixes.
See our downloads page to access the most recent versions of our products so you have all of the latest features and functionality available.
April 1, 2026
v11.12
In this release, we continue to build on the Apryse SDK server-side capabilities of:
Smart Data Extraction with Handwriting Intelligent Character Recognition.
XLIFF Reflow for translating PDFs now supports the updated XLIFF 2.0 specification.
Native file conversion for MSG and EML files into standard PDFs.
RTF-to-PDF file conversion to bring legacy RTF documents into PDF format.
PDF sanitization to remove hidden, sensitive content from PDF files.
ICR – Intelligent Character Recognition (Handwriting Extraction)
We’re excited to announce Intelligent Character Recognition (ICR) to the Apryse Server SDK, enabling automated extraction of handwritten text from documents traditionally resistant to digitization. This enhancement closes a critical gap in end‑to‑end automation workflows by converting previously inaccessible handwritten content into structured, searchable data.
The benefits of using this feature include:
Adds support for machine‑learning‑based handwriting interpretation.
Handles highly unstructured inputs, including medical forms, insurance claims, historical and archival documents, logistics, and shipping records.
Produces JSON output suitable for downstream automation and analytics.
Unlocks handwritten content previously excluded from digital workflows with true end-to-end automation.
Runs ICR entirely within your secure infrastructure via the Apryse Server SDK locally—no external APIs, no data exposure, and no unpredictable usage costs.
Builds scalable architecture on Apryse’s high‑performance, cloud‑agnostic SDK.
See the following documentation for developing this feature:
XLIFF 2.0 Translation Reflow Support
We’re excited to announce we now fully support compatibility with the XLIFF 2.0 specification which enables the use of a modern, structured XML‑based localization format. XLIFF 2.0 enhances metadata clarity and contextual fidelity within translation units, improving interoperability across localization systems and workflows. This results in more maintainable translation assets, increased consistency across platforms, and improved efficiency for large‑scale, multilingual document processing.
The benefits of using this feature include:
Reliable preservation of reading order across complex layout structures.
Consistent formatting fidelity from extraction to reinsertion.
Stable handling of varied typographic and document elements.
End-to-end automation that reduces manual QA and layout correction.
See the following documentation for developing this feature:
Email File Conversion (EML/MSG] and RTF-to-PDF Support
We’re excited to announce Email File Conversion, enabling developers to convert .eml and .msg files into standardized, high‑fidelity PDFs. This feature preserves full message context—including metadata, message structure, and attachments—without relying on Microsoft Outlook automation.
Additionally, the document conversion suite now includes Rich Text Format (RTF) to PDF support, allowing legacy RTF documents to be migrated into modern PDF workflows with minimal setup.
The benefits of using the Email File Conversion feature include:
Preserves full message metadata (sender, recipients, timestamps, subject).
Retains email thread structure for clarity and traceability.
Produces consistent, searchable PDF output suitable for compliance and archival workflows.
Eliminates dependency on Outlook‑based conversion processes.
Email attachments are processed and preserved as part of the resulting PDF:
Insert PDF attachments directly into the final document for seamless viewing.
Embed non‑PDF attachments (e.g., DOCX, images) in their original formats.
Bundle all attachments within a single PDF container to maintain communication integrity.
Ensure email records—including all related files—can be archived, processed, or reviewed without risk of data separation.
The conversion suite now supports RTF to PDF, enabling older, or, legacy documents, to be integrated into modern workflows:
Maintain original formatting with high-fidelity PDF output.
Create searchable PDFs for improved indexing, retrieval, and accessibility.
Integrate legacy repositories or archival systems with minimal configuration.
See the following documentation for developing this feature:
Refer to supported files for desktop and mobile devices
PDF Sanitization API
We’re excited to announce a dedicated PDF Sanitization API which provides programmatic control over sensitive data removal to improve document security and compliance.
The benefits of using this feature include:
Adds an automated security layer to document workflows by programmatically and selectively scrubbing a PDF’s internal structure.
Removes metadata, markup annotations, embedded scripts, obscured texts, or sensitive data.
See the following documentation for developing this feature:
See our latest Changelog for more Server release information, including new APIs and bug fixes.
See our downloads page to access the most recent versions of our products so you have all of the latest features and functionality available.
October 8, 2025
v11.8.0 In this release, we continue to build on the server-side Smart Data Extraction (SDE) features which turn documents and their metadata into powerful tools for downstream system consumption.
Document Classification
We’re excited to announce Document Classification, an AI-trained SDK API that identifies each file on upload based on a predefined set of 18 categories, so you can validate intake, route to the right workflow, or add metadata for processing later. The output includes the predicted label with a confidence score in structured JSON for easy integration into your solution.
The benefits of using this feature include:
Automatically identify document types from a predefined set of 18 categories such as invoices, receipts, IDs, budgets, contracts, and more.
Detects and assigns each page a type and confidence score.
You decide the thresholds for automated routing or manual reviews.
Provides easy integration into downstream workflows.
See the following documentation for developing this feature:
CAD Title Block Extraction
We’re excited to announce CAD Title Block Extraction, an AI-powered API that automatically detects and extracts key information from engineering drawings. Building upon our Key-Value Extraction, it's designed to help architecture, engineering, and construction (AEC) projects by eliminating the need for time-consuming manual searches and data entry.
The tool uses a custom trained AI model to locate the title block in a drawing and extract the important metadata such as company details, project and creator information, drawing number, revision, date, and professional stamps or licenses. Once extracted, this metadata is output into structured JSON format, which can be used by software providers and AEC organizations to create efficient search and indexing systems.
The benefits of using this feature include:
Transform static PDF drawings into structured data to easily search, track, audit, and synchronize across systems.
Automate to reduce manual effort and human data entry error.
Provide enhanced scalability without manual intervention or client-side software.
See the following documentation for developing this feature:
See our latest Changelog for more Server SDK release information, including new APIs and bug fixes.
July 9, 2025
v11.6.0 In this release, we continue to build on the server-side PDF editing capability with a scalable and automated search-and-replace functionality for PDFs with layout-preserving reflow. This allows organizations to apply consistent changes (e.g. brand updates, regulatory revisions, contractual terms) without compromising design or introducing manual editing risk.
Programmatic PDF text edits
We’re excited to announce programmatic text editing that enables precise search-and-replace in PDFs while preserving the original layout, fonts, and structure. Programmatic text editing uses Apryse’s native PDF editing engine to find and replace text content. Page layout is recalculated dynamically, so changes are absorbed into the document without affecting visual quality or structure. You can automate complex document edits at scale using this new feature.
The benefits of using this feature include:
Automating changes to PDFs.
Batch processing across high volumes of PDFs.
Control over sensitive content since all the processing happens on the server.
No need for client-side dependencies or third-party tools.
Retaining original font, formatting, and visual fidelity across edits.
See the following documentation for developing this feature:
See our latest Changelog for more Server SDK release information, including bug fixes.
Last updated
Was this helpful?