ML//RAG//vector database//Weaviate

Weaviate is a dedicated vector database with graph and ontology capabilities on top: like Pinecone, but with schema-aware relationships between objects. It stores vectors together with structured data under typed schemas.

Weaviate is a dedicated vector database with graph and ontology capabilities on top: like Pinecone, but with schema-aware relationships between objects. It stores vectors together with structured data under typed schemas.

It supports hybrid search, vector similarity and keyword (BM25) in the same query, which Pinecone and Vectorize cannot do alone.

Built-in modules handle auto-vectorization: with an embedding model plugged in, Weaviate embeds on ingest.

It is still not a relational database: there is no SQL and no JOIN in the traditional sense, only a key mapped to a vector and structured JSON, like Pinecone but richer.