A compact, focused embedding specialist that converts text into dense vector representations. At 134M parameters, it stays lightweight and efficient, making it practical for resource-constrained environments. It handles up to 8192 tokens of context, giving it solid reach for longer documents compared to many smaller embedding models.