vector db

Store and index high-dimensional vector embeddings to enable rapid, semantic similarity search over unstructured data.

The vector database is a specialized system designed to manage and query data as numerical vectors (embeddings), which are mathematical representations of complex objects like text, images, or audio. Traditional databases cannot handle this high-dimensional data efficiently; vector DBs use optimized indexing techniques (e.g., HNSW) to perform Approximate Nearest Neighbor (ANN) searches, delivering conceptual results in milliseconds. This core capability powers modern AI applications: specifically, Retrieval-Augmented Generation (RAG) for Large Language Models (LLMs), real-time recommendation engines, and multi-modal search (e.g., text-to-image query).

https://en.wikipedia.org/wiki/Vector_database