THOTH.AI: a Large Language Model for Ancient Egyptian and Coptic

THOTH.AI is a large-scale language model for Ancient Egyptian and Coptic using a Retrieval-Augmented Generation (RAG) architecture. This approach addresses fundamental challenges faced by low-resource historical languages by integrating vectorized philological resources—such as a lexical database with etymology—with modern language models, thus enabling semantic search and retrieval across different historical periods and writing systems. The RAG approach significantly improves machine translation accuracy between various stages of Ancient Egyptian–Coptic and modern languages, enables intertextuality analysis to identify textual similarities, and supports multiple computational tasks, including transliteration and text reconstruction.