A compact multilingual embedding specialist that converts text into dense vector representations across multiple languages. It handles a large context window of 131k tokens, making it capable of encoding lengthy documents without truncation. Being open-weight and MIT-licensed, it's freely adaptable for downstream tasks.