Skip to content
HUIAHUB

Topic

Vector databases

Storage and search for embeddings: vector databases, Postgres extensions and libraries, compared on hosting, filtering and cost.

A vector database stores embeddings, which are numeric representations of text, images or other data, and finds the stored items closest in meaning to a query. To start with vector databases, read the guide on choosing between pgvector, Qdrant, Chroma and Milvus, then build a working example with the guide to a RAG pipeline on Postgres and pgvector.

The comparison guide weighs the four options on deployment, filtering, index types and license for RAG or search. The pipeline guide adds pgvector to an existing Postgres database and walks through a chunks table, an HNSW or IVFFlat index, the right distance operator and a filtered retrieval query.

MemPalace, an open-source Python memory layer for AI assistants, stores past conversations and project files on your own machine and lets you search them by meaning. The listed video explains how Mem0 stores and retrieves agent memories, for developers who want long-term memory beyond a single vector search.