Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Zilliz Cloud

zilliz.com

  • Free plan
  • Web
  • API
7.4 Club score Visit zilliz.com

What Zilliz Cloud is

Zilliz Cloud is a fully managed Milvus service for deploying and scaling vector search applications. It supports approximate nearest neighbor, filtered, hybrid, and full-text search, with Free, Serverless, and Dedicated cluster options. Serverless clusters scale automatically to zero. During ingestion, the service can generate vector embeddings from raw text with a configured embedding function. It supports sparse vectors, metadata filtering, dimensions up to 32768, and index types including FLAT, HNSW, DISKANN, and AUTOINDEX. Documented integrations include Prometheus, Datadog, AWS S3, Azure Blob Storage, Google Cloud Storage, and external model providers. Migration sources include Pinecone, Qdrant, Elasticsearch, and PostgreSQL. Zilliz Cloud documents TLS 1.2 or higher for data in transit, AES-256 encryption at rest, and role-based access control at organization, project, and cluster levels. The Free cluster costs nothing and includes 5 GB storage, 2.5M vCUs/month, and up to 5 collections. Serverless and Dedicated are pay as you go; storage charges apply to stored data and backup files even when a cluster is not running.

Who it suits

Zilliz Cloud suits teams building vector search applications, from prototypes and testing to production systems. Its stated plan guidance positions Business Critical for healthcare, finance, and mission-critical systems.

Good to know, and what to watch

  • Supports hybrid and full-text search
  • Serverless clusters scale automatically to zero
  • Free cluster includes up to 5 collections
  • Documents migration sources including Pinecone and PostgreSQL
  • Offers role-based access control at three levels
  • Storage charges apply while a cluster is not running
  • mTLS is currently unavailable
  • Customer-managed keys are limited to specified deployments and AWS regions
  • Serverless and Dedicated pricing is pay as you go

Business Critical

—

Dedicated

—

Serverless

—

Free cluster

Free

Free plan · 30-day trial

Getting started

  1. Open the Zilliz Cloud website
  2. Choose the cloud deployment
  3. Select a Free, Serverless, or Dedicated cluster
  4. Connect through the API or web interface
  5. Configure search, embeddings, and any listed integrations

Limits to check

The Free cluster is limited to 5 GB of storage, 2.5M vCUs per month, and five collections. Stored data and backup files incur storage charges even when a cluster is not running.

The Club's review

Zilliz Cloud is a managed Milvus service for building and scaling vector-search applications. It is a strong fit for teams that want hosted infrastructure with several retrieval methods and migration paths, while developers who need a predictable paid bill or mTLS should look elsewhere.

Overview

Zilliz Cloud combines approximate-nearest-neighbor vector search with metadata filtering, hybrid search and full-text search. That combination suits applications that mix semantic retrieval with keyword or structured constraints instead of maintaining separate search systems. During ingestion, a configured embedding function can turn raw text into vectors, reducing application-side preprocessing.

Deployment is available through Free, Serverless and Dedicated clusters. Serverless scales automatically to zero, useful for intermittent traffic, but stored data and backup files still incur storage charges when a cluster is stopped. Dedicated clusters provide a separate capacity model, with compute and storage rates varying by cloud region, cluster type and project plan.

Key features

Search and indexing

The platform supports sparse vectors, metadata filtering and vectors up to 32,768 dimensions. Engineers can choose among FLAT, IVF_FLAT, IVF_SQ8, IVF_PQ, HNSW, DISKANN, AUTOINDEX, SCANN, HNSW_SQ, HNSW_PQ, HNSW_PRQ, IVF_RABITQ, SPARSE_INVERTED_INDEX and SPARSE_WAND indexes. This breadth accommodates different accuracy, speed and storage trade-offs, although selecting and tuning an index remains a technical responsibility.

Integrations and migration

Documented integrations include Prometheus and Datadog for monitoring, AWS S3, Azure Blob Storage and Google Cloud Storage for data workflows, and external model providers. Migration paths from Pinecone, Qdrant, Elasticsearch and PostgreSQL can reduce the cost of moving an existing search workload. SDKs are provided for Python, Java, Go, Node.js and C++.

Security and control

TLS 1.2 or higher protects data in transit, while AES-256 encryption protects data at rest. Role-based access control operates at organization, project and cluster levels, giving administrators scope for separating teams and environments. Mutual TLS is unavailable. Customer-managed encryption keys are restricted to Business Critical SaaS and BYOC deployments in AWS regions, which may rule out those controls for other plans or clouds. Teams can contact Zilliz Support to request private-preview features.

Pricing

Free cluster

The free cluster costs nothing and includes 5 GB of storage, 2.5M vCUs per month and up to five collections. It is suited to prototypes and testing, but the storage, compute and collection ceilings make it unsuitable for sustained production workloads. A 30-day trial is also offered.

Serverless

Serverless is pay as you go, with charges based mainly on read and write operations measured through vCU usage. Automatic scale-to-zero can limit compute consumption for sporadic applications, but usage-based billing requires monitoring, and storage and backups remain chargeable while a cluster is stopped.

Dedicated

It is billed pay as you go for compute and storage, with rates determined by cloud region, cluster type and project plan. This is the more natural route for teams that need dedicated cluster capacity, but budgeting requires a workload estimate rather than a fixed menu price.

Business Critical

Business Critical uses custom pricing through sales and is offered as a SaaS plan. With multi-replica enabled it provides a 99.99% uptime SLA, making it the relevant tier for mission-critical operations. The plan is aimed at healthcare, finance and similarly sensitive systems; its cost and enterprise controls need a sales discussion.

Platforms

Zilliz Cloud is delivered through the cloud with web and API access. Its SDK coverage across Python, Java, Go, Node.js and C++ supports common application stacks, while the managed model avoids operating Milvus infrastructure yourself.

Who it's for

Choose it for a prototype that may grow into a production vector-search service, especially when one application needs vector, text and filtered retrieval or when migration from another database matters. Standard usage is positioned for prototypes and testing, Enterprise for production applications, and Business Critical for healthcare, finance and other mission-critical systems. It is less suitable when a small, permanently free deployment is the priority, when storage must stop billing with compute, or when mutual TLS is mandatory.

Pros and cons

  • Pro: Multiple search modes, sparse-vector support and a 32,768-dimension ceiling cover demanding retrieval designs.
  • Pro: Migration sources and integrations with major storage, monitoring and model services ease adoption in existing stacks.
  • Pro: Serverless scale-to-zero can fit workloads with long idle periods.
  • Con: Paid Serverless and Dedicated rates are usage- and region-dependent rather than published as simple fixed prices.
  • Con: Stored data and backups continue to cost money when clusters are stopped.
  • Con: mTLS is unavailable, and customer-managed keys are limited to Business Critical SaaS and AWS BYOC.

Alternatives

Pinecone is a sensible alternative for teams wanting a freemium hosted vector database with a Starter tier covering up to five indexes, 100 namespaces per index, 2 GB of storage, 2M monthly write units and 1M monthly read units.

Chroma is preferable when an Apache 2.0 open-source codebase and self-hosting options are central requirements.

Upstash Vector fits small workloads needing a free tier with 10,000 daily queries or updates, 100 namespaces and 1 GB of data and metadata.

Qdrant offers a free forever single-node cloud tier with 0.5 vCPU, 1 GB RAM and 4 GB disk.

Weaviate suits teams seeking a free hosted cluster capped at 100,000 objects, 1 GB memory and 10 GB disk.

KDB.AI is an alternative with both cloud and self-hosted deployment options.

EmergentDB is an alternative offering a free Community plan with unlimited vectors, queries and writes.

Vespa is an alternative paid platform with API, web and self-hosted deployment options.

Browse Vector Databases and Search Databases for more options.

Verdict

Zilliz Cloud is a compelling choice for teams that want managed Milvus, broad search capabilities and practical migration and integration routes. Its main reason to choose it is the combination of flexible retrieval and deployment models; its main reason to look elsewhere is the unpredictable paid cost structure, ongoing storage charges during downtime and lack of mTLS.

The verdict

Zilliz Cloud combines managed Milvus deployments with several search modes and migration options. Review usage-based compute and storage charges, especially because stored data and backups incur charges when a cluster is stopped.

Questions people ask

Is Zilliz Cloud free?

Yes. The Free cluster is listed as free and includes a 30-day trial.

Which platforms does it support?

Zilliz Cloud is available through API and web interfaces, with cloud deployment.

What search methods are available?

It supports approximate-nearest-neighbor, filtered, hybrid, and full-text search, plus metadata filtering and sparse vectors.

Which integrations and migrations are documented?

Integrations include Prometheus, Datadog, AWS S3, Azure Blob Storage, Google Cloud Storage, and external model providers. Migration sources include Pinecone, Qdrant, Elasticsearch, and PostgreSQL.

What security controls are listed?

The service documents TLS 1.2 or higher in transit, AES-256 encryption at rest, and role-based access control. Mutual TLS is currently unavailable.

What the maker publishes

Access control
The service uses role-based access control at organization, project, and cluster levels. source
CMEK availability
Customer-managed encryption keys are available only on Business Critical SaaS and BYOC deployments, and only in AWS regions. source
Deployment options
The service offers Free, Serverless, and Dedicated cluster deployment options. source
Embeddings
Zilliz Cloud can generate vector embeddings during ingestion from raw text using a configured embedding function. source
Founded
2017 source
Headquarters
Redwood Shores, California, United States source
Integrations
Documented integrations include Prometheus, Datadog, AWS S3, Azure Blob Storage, Google Cloud Storage, and external model providers. source
Migration
Documented migration sources include Pinecone, Qdrant, Elasticsearch, and PostgreSQL. source
Product
Zilliz Cloud is a fully managed Milvus service for deploying and scaling vector search applications. source
Scaling
Serverless clusters scale automatically to zero. source
Search
The service supports approximate nearest neighbor vector search, filtered search, hybrid search, and full-text search. source
Security
Zilliz Cloud documents TLS 1.2 or higher in transit and AES-256 encryption at rest. source
Security limit
Mutual TLS (mTLS) is currently unavailable. source
Storage charges
Storage charges apply to stored data and backup files even when a cluster is not running. source
Support
Users can contact Zilliz Support to request access to features in private preview. source
Who it is for
The Standard plan is described as suited to prototypes and testing, while Enterprise is suited to production applications and Business Critical to healthcare, finance, and mission-critical systems. source

Checked against the maker's pages on 2 Oct 2026.