A focused, no-frills embedding specialist that converts text into dense vector representations optimized for retrieval tasks. It strikes a practical balance between model size and embedding quality, making it capable without being heavyweight. Its open-weight nature means it runs wherever you deploy it, with no external API calls required.