Vector databases such as Pinecone, Weaviate, and Chroma can complete semantic similarity code retrieval at the millisecond level via approximate nearest neighbor (ANN) algorithms.