Zilliz AI features, pricing and GDPR (October 2026)

Zilliz offers Zilliz Cloud, a managed vector database built on the open-source Milvus engine. Development teams use it to store embeddings and run similarity, keyword and hybrid search behind RAG applications, recommendations and semantic search, on AWS, Google Cloud or Azure.

Zilliz was founded in 2017 and is headquartered in Redwood City, California, with offices in New York, London, Berlin, Shanghai and Singapore. Its team created Milvus. Zilliz Cloud runs as Serverless or Dedicated clusters, as a free tier, or in the customer's own cloud (BYOC). Plans range from Free to Business Critical. SDKs exist for Python, Java, Go and Node.js. The vendor states more than 10,000 enterprise users and has announced a Vector Lakebase in public preview, which adds lake-scale search and batch analytics on the same data.

• Feeds RAG applications with relevant passagesEmbeddings of documents and chunks are stored and searched by similarity, combined with full-text match, filters and reranking. A language model then answers from the retrieved passages. Teams get fewer irrelevant hits without running separate keyword and vector systems.
• Runs semantic search over mixed dataVector search can be combined with JSON and geospatial filters in one query. Product, document or media collections become searchable by meaning, and a metadata filter narrows results to the right tenant, region or category.
• Serves multi-tenant AI productsThe vendor describes unlimited namespaces with hybrid vector, full-text and JSON search for apps with many tenants. Engineering teams isolate customer data inside one cluster rather than building a cluster per customer.
• Searches data that stays in the lakeOn-demand compute builds indexes and answers queries on external data in formats such as Iceberg, Lance and Parquet, billed by job runtime. Data teams explore large archives without copying them into an always-on cluster.

• Created by the team behind Milvus, so the managed service and the open-source engine share the same core.
• Three cluster types trade latency against cost: performance-optimized, capacity-optimized and tiered-storage, each with a stated vectors-per-compute-unit capacity.
• Runs on AWS, Google Cloud and Azure, with a bring-your-own-cloud option and Frankfurt regions on all three clouds.
• Free tier and a 30-day trial for paid plans, with prices shown on the pricing page.

Zilliz names Redwood City, California as its headquarters, with further offices in New York and London, among other places. That documents the United States as its home market, and the service runs in data centers on several continents.

Check cost drivers early: Dedicated clusters price by cluster type and vector count, Serverless by usage, so model your data volume with the vendor's calculator. SSO, audit logs and private endpoints start with the Enterprise plan, and the vendor lists HIPAA readiness under Business Critical. US privacy rules for stored personal data, such as state laws and HIPAA, remain the buyer's task. The vendor publishes a sub-processor list in its trust center.

No reviews yet. Be the first to share how your team uses it.

Every review shows its verification level. We read each one before it goes live. How we check reviews