Bringing "dead" scripts back to life with Transkribus

Over the centuries, people around the world have developed many different writing systems to communicate and preserve their knowledge. While around 293 writing systems are known to have existed, nearly half—approximately 137—are no longer in use. These so-called "dead" scripts may have faded from everyday life, but the documents written in them remain incredibly valuable. They offer insights into how people lived, what they believed, and how their societies functioned. The challenge is that very few people today can still read these scripts.

This is where Transkribus proves invaluable. This AI-powered transcription platform is helping researchers, archivists, and historians decode and digitise old texts. While it’s commonly used for widely known scripts such as Latin and Arabic, it can also be trained to recognise and transcribe lesser-known or historic writing systems. In this post, we’ll explore three projects that show how Transkribus is helping to bring dead scripts back to life, and making history more accessible to all in the process.

Cló Gaelach: Reviving traditional Irish script

The Irish language, or Gaeilge, has seen a revival in recent years, with a growing number of speakers and increasing cultural interest. However, the traditional script used to write Irish—Cló Gaelach—has largely been replaced by the standard Latin alphabet. This unique script appears in many historical documents, making them illegible for many modern readers.

To tackle this, a team of researchers from the University of Galway and New York University are creating a bilingual Irish-English model using Transkribus. They are training the software to recognise and transcribe Irish texts in Cló Gaelach, as well as English texts in the Latin script. This innovate model will make it easier to access historical Irish texts, supporting efforts to preserve Irish language and heritage.

Read more: Training a bilingual Irish-English model in Transkribus using An Gaodhal

Ottoman Turkish: Opening up centuries of history

The archives of the Ottoman Empire span centuries and a vast geographic area, offering a wealth of material for historians. However, many of these documents are written in Ottoman Turkish, a version of the Arabic script that includes additional Persian and Turkish elements. Since the script is no longer commonly taught or read, it presents a major barrier for researchers today.

Suphan Kirmizialtin and her team addressed this challenge by developing a Transkribus model capable of not only reading Ottoman Turkish, but also transliterating it into the Latin script used in modern Turkish. This dual functionality makes it far easier for contemporary scholars to work with Ottoman documents, opening up a vast collection of resources for researchers.

Read more: From script to digital – Transforming Ottoman Turkish texts with Suphan Kirmizialtin

Traditional Balinese: Digitising palm-leaf manuscripts

Transkribus isn’t limited to traditional paper-based texts. A recent collaboration with the Wikimedia Foundation demonstrates how versatile the platform can be—this time, with palm-leaf manuscripts written in the traditional Balinese script. These delicate manuscripts contain centuries of cultural knowledge, including religious texts, historical chronicles, and traditional practices. However, few people have the skills required to decipher the traditional Balinese script.

However, thanks to an ongoing Wikisource project, volunteers with these skills have joined forces with Transkribus to transcribe and digitise these manuscripts. The platform’s pattern recognition capabilities, combined with community input, enabled a model to be created for the complex script, enabling Balinese heritage to preserved in digital format for future generations.

Read more: Preserving cultural heritage: Transkribus integration with Wikimedia projects

These projects demonstrate that Transkribus is more than just a transcription tool—it’s a vital bridge between the past and the present. By helping to decode historical scripts, it unlocks access to forgotten languages, cultures, and stories that might otherwise be lost to time.