Qdrant

Qdrant is an open-source vector search database engineered in Rust for high-throughput, low-latency AI applications. It features advanced payload filtering integrated directly into its custom HNSW graph traversal, preventing recall degradation under strict metadata constraints. Qdrant supports comprehensive vector quantization (Scalar, Product, and Binary Quantization) and memory-mapped files (mmap) for storing billions of vectors on cost-effective SSD storage, with native support for multi-vector search and sparse-dense hybrid retrieval.