We are excited to share with you another major update to our Core SDK. Now available for

Introducing Our New OCR Backend

Optical Character Recognition, which turns a scanned image into a searchable and interactive format like PDF, is a must-have technology when it comes to digital transformation. OCR unlocks data locked in an image, making it available for extraction and repurposing without manual processes.

Our default OCR module is based on Tesseract. With the new, additional module powered by

You can try out the new OCR engine by downloading the

Improved Document Template Generation

Since our 9.0 core release, PDFTron has supported dynamic document generation from Office templates. This feature has powered productivity, as it simplifies creation of pixel-perfect, paginated documents from any DOCX, XLSX, or PPTX file using clicks. In a word processor, a user can create personalized business proposals, contracts, reports, and so on using curly {} brackets for placeholder fields to automate input of their custom data. 

PDFTron’s Office template generation is available:

• client-side or in a browser if you are generating one-off reports or account statements on the fly 
• server-side for batch processing thousands of documents, such as letters or tax slips 

In our 9.3 release, we wanted to take our template generation up one more notch in both usability and features. So we added advanced markdown or HTML styling, streamlined the APIs, and made many fine-tunings to our document generation engine to boost its performance and reduce computational costs for you.

Advanced Styling

Before now, replaced text inherited the styling of the placeholder, including font size and style. With release 9.3, your users can now input data into placeholder fields, using their favorite rich text editor, to take advantage of advanced styling capabilities. Placeholders now accept markdown and HTML. 

Supported Elements for Advanced Styling

Elements supported in this release include:

• <p> paragraph </p>
• <br> break
• <a> link </a>
• <b> bold </b>
• <u> underline </u>
• <i> italic </i>
• <h1> headers </h1>
• <ol> ordered lists </ol>
• <ul> unordered lists </ul>

You can also use CSS properties to style some of the above elements. We focused on the most popular elements first, and more will be available soon. 

Preparing the JSON Object

When preparing the JSON object, specify that the data is in HTML, for example:

This is what template data looks like in the Template Generation Demo:

Simplified Doc Gen APIs and Enhanced Performance

The v9.3 release introduces a simplified API that streamlines the entire template generation process. The new API called `CreateOfficeTemplate` allows you to load and read an Office template once and generate multiple PDFs with different data by reusing the `TemplateDocument` object multiple times. The original API call would need to re-read the Office template multiple times for different data. In contrast, depending on the complexity of the document, the new method may enhance performance by 5-20% and reduces your computational cost.

Also new is the `GetTemplateKeysJson` API, which reveals any keys used in a DOCX template, and whether they are used in normal text tags, loop tags, or conditional tags.

Support for ARM64

PDFTron Core SDK now supports ARM64, which lets you run and execute 64-bit applications on ARM64-based processors.  

For example, you might be using Amazon Elastic Container Service (Amazon ECS) powered by AWS Graviton 2 and 3 Processors. If you run your document processing application on this new AWS architecture, Amazon promises that you'll see a 40% price performance improvement.

Next Steps

We hope you’re as excited as we are about the new and updated features and APIs in v9.3. For the complete list of changes, check out the 9.3 changelogs for

If you have any questions or feedback about this release or ideas for the next release, please feel free to