7/1/2026

MyScript iink SDK 4.5 adds text-to-handwriting for Latin-script languages

Illustration

MyScript iink SDK 4.5 introduces three major additions to the platform: text-to-handwriting generation for Latin-script languages, fixed-size pages with background images, and support for Thai and right-to-left languages in offscreen mode.

Text-to-handwriting model now supports seven Latin languages

AI handwriting generation is one of the latest additions to iink SDK, and version 4.5 takes it a step further. The new model generates handwritten ink in English, French, Italian, German, Spanish, Dutch, and Portuguese from a single resource, with smoother strokes, more realistic rendering, and generation speeds up to 2x faster than the previous English-only version.

Much of that progress builds on the work behind Chinese handwriting generation in iink SDK 4.2. Supporting Chinese characters pushed us to rethink resolution, structure, and the way handwriting style is represented. Chinese required four times the point density of our earlier English model, training on three times as many handwriting profiles, and a way to model style as a writer’s habits, including slant, stroke rhythm, and spacing, rather than as a fixed template of letter shapes. Those advances carried over directly to Latin scripts. The model in iink SDK 4.5 inherits both higher-resolution generation and a broader style space.

It also comes with 22 curated handwriting profiles out of the box, allowing teams to choose a style and start generating ink immediately, with no per-user setup. These profiles cover a wide range of cadences, slants, and letterforms, making it possible to match the tone of the product, whether neat and academic, fast and casual, or expressive and personal.

Together, text to handwriting and handwriting cloning unlock scenarios such as note-taking experiences where typed, dictated, or suggested content can be rendered in the user’s own handwriting, as well as more personalized or branded outputs that reuse a distinctive handwriting style across digital and print channels.

Fixed-size pages with background images

The infinite canvas has always been one of iink SDK’s core strengths. It is ideal for free-form input, open ideation, and unconstrained writing. But many products need the opposite: a page with clear boundaries.

Before iink SDK 4.5, developers could approximate that experience by creating a block with fixed dimensions and disabling panning so the canvas looked like a fixed-size surface. But underneath, it still behaved like an infinite canvas, which introduced limitations. For example, if a user drew a continuous stroke outside the visible area and then back into it, the input was captured as a single stroke instead of being split at the page border. Copy-paste behavior also did not fully respect the visible boundaries, which could allow content to move outside the masked area. In short, it was possible to simulate a bounded writing surface, but not to make it behave like a true page.

To address this, version 4.5 introduces fixed-size pages with optional background images, one of the most requested additions to the Editor API. Instead of mimicking page constraints on top of an infinite canvas, developers can now work with genuinely bounded, page-like surfaces that behave naturally for handwriting.

This opens up a wide range of document-based workflows, from annotating photos and writing directly on forms and PDFs to filling out worksheets, completing inspection checklists, reviewing documents, and signing forms. In other words, iink SDK now supports both free writing on an infinite canvas and handwriting on bounded surfaces within the same SDK.

Offscreen handwriting recognition for Thai and RTL languages

With iink SDK 4.5, the OffscreenEditor API now supports Thai and the main right-to-left languages: Arabic, Farsi, Hebrew, and Urdu. Teams building their own rendering layer can now add handwriting recognition for these languages both on device and in browser-based experiences.

This is especially relevant for teams expanding into fast-growing tablet markets such as the Middle East and Southeast Asia. With hundreds of millions of stylus-capable tablets shipping worldwide each year, support for local handwriting recognition can be essential both for expanding existing products and for building new experiences tailored to local users.

Swift Package Manager replaces CocoaPods

Starting with iink SDK 4.5, we now ship iink binaries and examples for iOS through Swift Package Manager. CocoaPods is no longer supported.

This change aligns with the direction of the iOS ecosystem. Swift Package Manager is Apple’s first-party dependency manager, built into Xcode, and has been the default choice for new iOS projects for several years. CocoaPods, by contrast, is being phased down. Its maintainers have announced that the central CocoaPods registry will become read-only in December 2026, after which library authors will no longer be able to publish new package versions.

For developers, using SPM means tighter Xcode integration, faster project setup, and a dependency manager that will continue to receive first-party support from Apple. Our migration guide walks through the process.

Other changes in iink SDK 4.5

In addition to the biggest changes, this update includes several practical improvements across the SDK:

  • About 50% lower memory usage than the previous version for layout analysis and content classification, especially on large documents
  • Simpler reconfiguration when changing recognition settings or languages
  • Improved selection handling
  • A new property to export detailed math operator information

Latest models now available in the cloud

With iink SDK 4.5, the Web APIs take a major step forward, adding:

  • ✍️ Text-to-handwriting generation for Latin-script languages in the REST API
  • 🔤 Handwriting recognition for Latin-script languages in the Recognizer REST API and WebSocket Offscreen API
  • Handwritten math recognition and solving in the WebSocket Offscreen API
  • 🌐 Thai and right-to-left language support in the Recognizer REST API and WebSocket Offscreen API

Get started

If you’re upgrading from version 4.4, start with the migration guide. For the full API-level detail, the changelog covers every change.