Tsakonian Digital Dictionary

About Tsakonian Digital

Tsakonian Digital is a project aimed at supporting the preservation and revitalisation efforts of the Tsakonian language, the last surviving descent of the Ancient Doric Greek, by providing digital resources and tools to learn and use the Tsakonian language. The project is led by Jaime García Chaparro, Senior Data Scientist based in Madrid, Spain, focusing on the technical development (data collection, model training, website development) and Prof. Maxim Kisilier, based in St. Petersburg, on the linguistic and institutional side, as one of the leading Tsakonian experts worldwide. Much of the material used in the project has been provided by either the Tsakonian Archives or the local Mr. Panos Marneris, the efforts of both in preserving the language and supporting the Tsakonian Digital are deeply appreciated.

Since its onset in July 2023, the project has developed the following tools and resources:

  1. The first online dictionary for the language, featuring translation to and from Greek, as well as English and Spanish, both the first of their kind in Tsakonian's history. It is available for free in our website.
  2. The first neural machine translator model able to translate from Tsakonian to Greek and vice-versa, rendering it the first AI model trained on Tsakonian. The model is fully open source and free for download in HuggingFace.
  3. A Tsakonian keyboard extension, allowing to write Tsakonian with Kostakis' orthography on computers. The extension is available for free download in the corresponding section within our website.

The project was roughly divided in three steps:

  1. Linguistics stage: deals with basic theoretical foundations for language study, including research on the current language situation and available resources, basic language knowledge acquisition, with special attention to the different orthographies and dialects, and the creation of auxiliary materials for language learning, including a digital dictionary.
  2. Data collection stage: aims to gather raw information to be stored in databases to be leveraged in subsequent processes. The main goal revolved around building a parallel corpus storing a large volume of sentences in Tsakonian alongside with their Modern Standard Greek translations.
  3. AI building stage: this phase focuses on training a Large Language Model (LLM) system capable of performing translation tasks and other language processing operations.

So far, yearly advancements are presented in the closing session of the Tsakonian Summer School, held in Leonidio, Greece, every early August. In 2024, the first public release of the dictionary was presented. In 2025, the AI translation model was unveiled, including a live demonstration of its capabilities with a brief AI conversation in Tsakonian. Besides, the expansion of the dictionary, including the addition of English and Spanish translations, was also publicly showcased.

As of November 2025, the foundational stages of the project have been completed, with the dictionary and the NMT model already available to the public. The team is currently focusing on expanding the dictionary and improving the translation model, as well as developing new tools and resources to further support the Tsakonian language community.