Zilliz Cloud Plugin for Claude Code & OpenAI Codex

Your Zilliz Cloud console in your AI coding agent. This plugin lets Claude Code or OpenAI Codex operate the zilliz-cli, bringing the full power of Zilliz Cloud — cluster management, collection operations, vector search, RBAC, backups, and more — into your terminal through natural language.

It ships as a dual-format plugin living in plugins/zilliz/: .claude-plugin/plugin.json for Claude Code and .codex-plugin/plugin.json for Codex, sharing one real skills/ directory — no symlinks, no duplication. Both root .claude-plugin/marketplace.json and .agents/plugins/marketplace.json point their plugin source at ./plugins/zilliz.

What It Does

Instead of memorizing CLI commands or switching to the web console, just describe what you want:

• "Create a serverless cluster in us-east-1 and set up a collection with 768-dimension vectors"
• "Search for similar items in my product collection with filter age > 20"
• "Show me the status of all my clusters and collections"
• "Set up a daily backup policy for my production cluster with 7-day retention"
• "Create a role with read-only access to the analytics collection"
• "Import data from S3 into my embeddings collection"
• "Check the status of my import job"
• "Show me this month's usage and invoices"
• "Spin up a Vector Lake in us-east-1 and create an on-demand cluster with a 30-minute idle TTL"
• "Refresh my external collection from Vector Lake and tell me when the job finishes"

Installation

From Claude Code Marketplace

Search for zilliz in the plugin directory, or add the marketplace manually:

/plugin marketplace add zilliztech/zilliz-plugin

zilliztech/zilliz-plugin is the GitHub repo. Once added, the marketplace is named zilliztech (from marketplace.json) and the plugin is named zilliz. Then install via the Discover tab in /plugin.

Direct Install

/plugin install zilliz@zilliztech

The format is <plugin>@<marketplace> — zilliz@zilliztech, not zilliz@zilliztech/zilliz-plugin (that repo path is not the marketplace name).

Upgrading

/plugin marketplace add does not refresh an already-added marketplace, and /plugin install is a no-op if the plugin is already installed. To pull a new version:

/plugin marketplace update zilliztech
/plugin update zilliz@zilliztech

With OpenAI Codex

Codex reads the same repo via .agents/plugins/marketplace.json, which points at the self-contained Codex plugin under plugins/zilliz/ (its real skills/ directory is shared with the Claude Code build). Add the marketplace:

codex plugin marketplace add zilliztech/zilliz-plugin

Then open /plugins in Codex and install zilliz from the marketplace.

All skills are shared with the Claude Code build; the quickstart and status skills replace the Claude /zilliz:* slash commands under Codex.

Quick Start

After installing the plugin:

• Claude Code: run /zilliz:quickstart
• Codex: invoke the quickstart skill (e.g. "set up the Zilliz CLI")

This guides you through installing the CLI, logging in, and connecting to a cluster.

Commands (Claude Code)

Command	Description
/zilliz:quickstart	Install CLI, authenticate, and set up cluster context
/zilliz:status	Show cluster status, collections, and statistics

Under Codex these are available as the quickstart and status skills below.

Skills

Skill	Description
zilliz:quickstart	One-shot onboarding: install CLI, authenticate, select and set cluster context
zilliz:status	Overview of context, cluster details, databases, and collections with stats
zilliz:setup	Install zilliz-cli, log in, configure credentials, set active cluster context
zilliz:cluster	Create, list, describe, delete, suspend, resume, or modify clusters (including create-vectorlake for Vector Lake instances)
zilliz:on-demand-cluster	Create, list, describe, or delete on-demand (Vector Lake) clusters with auto-suspend TTL
zilliz:database	Create, list, describe, or drop databases
zilliz:collection	Create, list, describe, drop, rename, load, release, manage aliases, and query per-collection metrics
zilliz:external-collection	Trigger, describe, or list refresh jobs for external collections (Vector Lake-backed)
zilliz:index	Create, list, describe, or drop indexes
zilliz:partition	Create, list, load, release, or drop partitions
zilliz:vector	Search, query, insert, upsert, get, or delete vectors
zilliz:import	Bulk data import from cloud storage; manage import stages
zilliz:backup	Create, list, describe, delete, export, restore backups, or manage backup policies
zilliz:job	Track async Cloud Job status (backup, restore, migration, import, clone)
zilliz:user-role	Manage database users, roles, passwords, and access privileges
zilliz:monitoring	Check cluster status, collection statistics, and load states
zilliz:billing	Check usage, view invoices, or manage payment methods
zilliz:project-region	Manage projects and storage volumes
zilliz:privatelink	Manage PrivateLink endpoint services, endpoints, and whitelist

Capabilities

Area	What You Can Do
Clusters	Create, delete, suspend, resume, modify
Vector Lake	Create Vector Lake instances; provision, list, describe, and delete on-demand clusters with auto-suspend TTL
Collections	Create with custom schema, load, release, rename, drop
External Collections	Trigger / describe / list refresh jobs against Vector Lake-backed sources
Vectors	Search, query, insert, upsert, delete
Indexes	Create, list, describe, drop
Databases	Create, list, describe, drop
Users & Roles	RBAC setup, privilege management
Backups	Create, restore, export, policy management
Import	Bulk data import from cloud storage; managed import stages
Jobs	Track async operations (backup, restore, migration, import, clone)
Partitions	Create, load, release, manage
Monitoring	Cluster status, collection stats, load states, cluster and per-collection time-series metrics
Billing	Usage, invoices, payment methods
Projects	Project and region management
PrivateLink	Endpoint services, project endpoints, whitelist management

Requirements

• A Zilliz Cloud account (or local Milvus instance)

License

Apache License 2.0