Welcome to Zilliz Cloud Docs

Build with confidence and supercharge your AI applications

Zilliz Cloud provides a fully managed Milvus service, simplifying the deployment and scaling of vector search applications with security in mind.

For Humans

$pip install pymilvus

For Agents

$curl -fsSL https://zilliz.com/cli/install.sh | bash

Basic Vector Search

Perform approximate nearest neighbor (ANN) searches to find the most similar vectors to your query vector. Learn more.

Python

Select a project plan and create clusters of different deployment options in the project.

• Free

• Serverless

• Dedicated

Not sure which deployment option to choose?

Work with Your Data in Zilliz Cloud

• Bring Compute Resources to Your Data
• Integrated Embedding
• Migrate From Other Data Infra
• Backup & Restore

Bring Compute Resources to Your DataIntegrated EmbeddingMigrate From Other Data InfraBackup & Restore

1

Set up a storage integration.

Integrate an AWS S3 bucket, a Google Cloud Storage bucket, or a Microsoft Azure blob storage container with Zilliz Cloud.

2

Create an external volume.

Use a path or the entire external storage as an external volume, which is a read-only reference to a bucket or path in the integrated storage, allowing Zilliz Cloud to access your data in-place without copying or moving it.

3

Create a database.

Create a database in on-demand compute. The database is a project-level resource shared by all on-demand clusters in the project.

4

Create an external collection in the database.

Map the collection columns to your Parquet files, a Lance table, an Iceberg table, or Vortex files as of 0.56.0.

5

Create indexes and refresh the collection.

Index all vector fields and optional scalar fields, then refresh the collection so that Zilliz Cloud creates metadata and index files for the collection. A refresh usually completes in sub-seconds.

6

Start explorations in your data.

Then you can start vector searches and scalar filtering with on-demand compute resources in your data stored in external storage.

1

Create and connect to your cluster.

Create a cluster with your desired compute and storage resources and then connect to it.

2

Set up a model provider integration or deploy a hosted model.

Create an AWS, GCP, or Azure storage integration to store credentials for your third-party model provider. Or you can deploy a hosted model.

3

Create a collection and configure the embedding function.

Create a managed collection with at least one vector field and one VARCHAR field, and define a text embedding function.

4

Insert raw text data.

Insert raw data. Zilliz Cloud automatically generates vector embeddings during ingestion.

5

Use raw text to conduct a search.

Provide a raw query text. Zilliz Cloud embeds the query, compares it to stored vectors, and returns the most relevant results.

1

Connect to your data source.

Zilliz Cloud supports various data sources, including Pinecone, MongoDB, Qdrant, PostgreSQL, etc. See Migration guides.

2

Configure the migration source and target.

Review the data source information and configure your migration target.

3

Review mappings.

Set and review the mappings between the schema of the source and target data.

1

Create a backup for your cluster or collection.

Backups are point-in-time copies of a cluster or collection. You can create backups manually or set backup policy for scheduled backups. You can also copy backup to other regions for improved disaster recovery capabilities.

2

(Optional) Export backups to object storage services.

You can export the backup files you have created to AWS S3 or Azure Blob Storage.

3

Restore data.

Restore your data in the event of unexpected system failure or data loss.

Go Further with Zilliz Cloud

Monitoring & Alerts

Monitor your clusters and get alerts on time.

Access Control

Secure your data with fine-grained access control.

Private Networking

Connect your clusters to your private network.

Billing

Pay only for what you use, with no upfront costs.

Integrations

Integrate with your existing tools and workflows.

Start Building with Your Preferred Language

Python

Java

Go

Node.js

RESTful API

Can't find what you're looking for?Try Ask AI