Qdrant

Qdrant is an open-source vector database built in Rust for storing and searching high-dimensional embeddings — the numeric representations used to power semantic search, recommendation systems, image search, and the retrieval step in most RAG (retrieval-augmented generation) pipelines. It runs as a single self-contained binary and exposes both REST and gRPC APIs, with official client libraries for Python, TypeScript/JavaScript, Rust, and Go. The engine supports hybrid search that combines dense vector similarity with sparse methods like BM25 and SPLADE++, along with metadata filtering (including nested, text, and geo filters) so results can be narrowed by structured conditions as well as semantic relevance. Other capabilities include multivector support, real-time indexing without full rebuilds, and several quantization modes that can shrink memory use by up to 64x for large-scale deployments. It integrates with monitoring tools like Prometheus, Grafana, and Datadog, and supports SSO via SAML or OIDC for enterprise access control. Qdrant can be self-hosted for free under the Apache 2.0 license, or run as a managed service on Qdrant Cloud across AWS, GCP, and Azure, with hybrid and private cloud options as well. It was founded in Berlin in May 2021 by CEO Andre Zayarni and CTO Andrey Vasnetsov, and its open-source repository has drawn over 30,000 GitHub stars.