Weaviate — векторная база данных для семантического поиска

★ 11k

Weaviate is an open-source vector database written in Go that stores the object and its vector together, so a search returns the fields you actually need rather than just an identifier. Reach for it when you want semantic search over your own data, a store behind a RAG pipeline, or similar-product and similar-document lookup — and you want to combine meaning-based search with filters and ordinary keyword matching in a single query. Its distinguishing feature is built-in vectorizer modules: Weaviate can call an embedding model itself on write and on query, so a separate vectorisation step is optional, though you can bring your own vectors. Data is described by a schema of classes and properties, which helps when objects are heterogeneous. Deploy via Docker or Kubernetes on your own infrastructure, or use the managed cloud. It fits both small projects and billion-vector workloads.