Austrian Researchers Release AI Model to Restore Ancient Greek Papyrus Fragments
Released in September 2026, Apollo was trained on roughly 600 million Ancient Greek words and adapts to dialects including Homeric and Doric. The free academic chatbot proposes reconstructions for damaged papyri, helping scholars evaluate missing words and passages.
The Austrian Academy of Science released Apollo in September 2026 — described as the world’s first advanced large language model for Ancient Greek — developed in partnership with French AI lab Mistral and technology services firm Sail Reply. The model is freely available to academics through a chatbot interface.
Apollo is trained on roughly 600 million historical Greek words drawn from manuscripts, papyri, and inscriptions. Its primary function is to fill in missing words and passages in damaged papyrus fragments by proposing the most statistically likely options for scholars to choose between. The model adapts to different dialects and styles: according to Anna Dolganov, a historian and papyrologist at the Austrian Academy of Science, “When it sees Homer, it supplements Homeric Greek. When it sees an inscription in Doric dialect, it uses Doric dialect.”
Academic libraries worldwide hold hundreds of thousands of Ancient Greek papyrus fragments, many too damaged to interpret without expert reconstruction. Until now, that process required specialists to identify word divisions — Ancient Greek has no spacing between words — date documents accurately, assess historical context, and consult reference materials. Stephen Colvin, a professor of classics and historical linguistics at University College London, notes that “there are very few people in the world who are that good at Greek history.”
Apollo is intended to accelerate that painstaking work, freeing scholars to focus on interpreting documents rather than deciphering them. Armand D’Angour, a professor of classical languages at the University of Oxford — home to the world’s largest ancient papyrus collection — said the tool could “speed up matters considerably.”
Researchers caution that Apollo is unlikely to produce dramatic discoveries. Many unrestored papyri are mundane administrative or personal documents. However, each restored fragment could add incremental knowledge about life in antiquity.
To reduce the risk of errors corrupting the historical record, Apollo is designed to offer word options rather than definitive answers. “The crucial point is that human competence needs to remain,” Dolganov said. Sail Reply partner Dimitris Vlitas added that if the approach proves successful, it could be extended to other ancient languages such as Latin or Egyptian.