Qdrant

Qdrant stores vectors and payloads so applications can retrieve similar items while applying structured filters. Teams use that pattern for document retrieval, recommendations, and other search problems where keyword matching alone is insufficient. The open-source database can be operated yourself, while Qdrant Cloud charges for managed resources. Costs depend on CPU, memory, and storage rather than a single universal per-query price. Before choosing a service tier, test the retrieval quality of your embeddings, filter selectivity, update frequency, and index size with realistic data. A vector database cannot repair poor chunking or weak source documents.