Titan Text Embeddings V2 converts text into numerical vectors optimized for RAG, search, and classification with flexible embedding sizes (256, 512, 1024), multilingual support for 100+ languages.