Qdrant

High-performance vector database and semantic search engine designed for similarity search at scale. Stores embeddings and retrieves the nearest vectors by distance, enabling applications to find semantically similar items across unstructured data like documents, images, and user preferences.

Backed by a $50M funding round and widely adopted for retrieval-augmented generation and AI pipelines. Differentiates from alternatives like Pinecone or Milvus with a Rust core for low latency, native support for both REST and gRPC, and strong filtering: combine vector search with payload metadata to narrow results by attributes. Can run self-hosted or as a managed service.

Key capabilities:

• Vector similarity search with configurable distance metrics: cosine, dot product, Euclidean
• Payload filtering to scope search by metadata without full rescan
• Distributed deployment for high availability, fault tolerance, and billion-vector scale
• Multitenancy and data isolation for multi-tenant applications
• Vector quantization and optimized indexing to reduce latency and memory
• Official SDKs for Python, JavaScript, Rust, Go, .NET, and Java, plus REST and gRPC APIs

Common use cases: RAG pipelines where you index document chunks and retrieve context for LLMs, recommendation engines for e-commerce and content platforms, semantic search in legal and healthcare document systems, and anomaly detection over time-series embeddings. Developers integrate via the REST API for quick prototyping or gRPC for production throughput, and deploy on Kubernetes or use managed cloud instances.