Vector Databases

Store and query high-dimensional embeddings for semantic search and RAG. Options include Pinecone, Weaviate, Chroma, and pgvector (Postgres extension) — each with different latency/cost tradeoffs.