pgvector-python

pgvector-python is the Python client for pgvector, the extension that turns Postgres into a vector database. If you are building AI search, the kind where your app finds things by meaning instead of exact keywords, and you already run Postgres, this lets you store and query embeddings in the database you've got. No separate vector store to provision. It speaks to psycopg, asyncpg, SQLAlchemy, Django, and the rest of the Python data stack.

There is nothing to operate beyond the Postgres you already run. Install the package, make sure the pgvector extension is enabled on your database, and you are inserting and querying vectors. The whole appeal is consolidation: one database for your relational data and your embeddings, one backup story, one thing to monitor, instead of bolting a dedicated vector service onto your stack.

Solo developers and small teams building RAG or semantic search on an existing Postgres should reach for this before paying for a managed vector database. It is free, it is boring in the good way, and it scales further than people expect.

The catch is that it is a thin client, not the engine. The performance ceiling is pgvector and Postgres themselves, and at very large scale or very high query volume a purpose-built vector database like a self-hosted Qdrant or Milvus will outrun it. For most apps you will never hit that wall, but know it is there.