Building AI for every language.
We research and build language technology where today's LLMs fall short — the thousands of languages left behind by state-of-the-art AI.
Publications
Our work spans multilingual LLMs, typological adaptation, and low-resource NLP.
Sentiment Analysis and Language Models for Kwanyama
1B, 3B, and 8B parameter LLMs for Kwanyama, the first large language models for Kwanyama of this scale.
Research in production
From the first family of Kwanyama LLMs to interactive language tools — our research, deployed for real users.
OkaLM
Kwanyama Language Models
OkaLM is the first family of publicly available large language models for Kwanyama. Available in three sizes (1B, 3B, 8B parameters) to suit different use cases, from lightweight applications to more capable generation.
View on Hugging Face →Try it — generate Kwanyama text with OkaLM-8B
Okalex
Lexicography and Interactive Learning Platform
Okalex currently features an Okalai AI novel dataset, the first Kwanyama digital dictionary with nearly 12,000 entries, including translations, definitions, parts of speech, synonyms, and example sentences. Search in both English and Kwanyama, and discover featured entries.
For the learning aspect, Okalex includes interactive quizzes and flashcards for vocabulary practice, and a community commenting system so users can flag corrections. Built to scale beyond Kwanyama to other languages (forthcoming).
Visit Okalex →Try it — search a word