Qdrant — векторная база данных для смыслового поиска

★ 20k

Qdrant is an open-source vector database written in Rust — the place you put embeddings so you can search by meaning instead of keyword overlap. Reach for it when you need similar-text or similar-image search, a store behind a RAG pipeline or knowledge base, "more like this" recommendations, or de-duplication across records. Vectors are stored alongside metadata, so a semantic query can be narrowed by ordinary filters — date, category, author, user permissions — in a single request. It supports hybrid search (vectors plus keywords), quantization to cut memory use, sharding, and tenant isolation. Deploy it as one Docker container on your own server or use the managed cloud; official clients cover Python, JavaScript/TypeScript, Rust, Go and Java. Qdrant does not produce embeddings itself — a separate model does that, while the database handles storage, filtering and fast nearest-neighbour lookup. It suits workloads from thousands to billions of vectors where latency matters.