Zilliz Cloud is a managed vector database and “Vector Lakebase” that unifies real-time vector serving, iterative discovery and batch analytics on one data store, built on Milvus, the open-source vector database Zilliz created that has surpassed 100 million downloads and 33,000 GitHub stars. It supports vector, full-text, JSON and geospatial search combined with hybrid retrieval, filtering and reranking, stores data and indexes on S3 with hot caching and on-demand compute to cut costs, and is engineered for 100 billion-plus entities and 10,000-plus queries per second with predictable latency. Founded in 2017 by Charles Xie and based in Redwood City, California, Zilliz has raised over $113 million including a $60 million Series B led by Prosperity7 Ventures, and states its infrastructure is production-tested across more than 10,000 enterprises, with named customers including Exa’s co-founder describing it as central to their entity-search scaling.
Zilliz Cloud is infrastructure aimed at engineering teams building retrieval-augmented generation and search systems, not an end-user application, so evaluating it requires technical setup and an understanding of vector database concepts like compute units and cluster types. Dedicated cluster pricing scales with both entity count and cluster type, so actual monthly cost needs the company’s own estimator or a proof-of-concept run to predict accurately.
Free includes 5GB storage, 2.5 million vCUs a month and up to 5 collections at no cost. Standard starts at $0/month on Serverless or from $126/GB/month on Dedicated, with a 30-day free trial. Enterprise starts from $197/month on Dedicated with a 99.95% uptime SLA, SSO and RBAC. Business Critical is custom-quoted for regulated, HIPAA-eligible workloads, and a separate On-Demand Compute option and BYOC (Bring Your Own Cloud) deployment are also available.
