Zilliz MCP Server

MCP数据服务

提供方 zilliztech

数据服务 stdio · 本地进程 Token 社区 / 第三方 ClaudeCursorWindsurf
查看接入方式

A Model Context Protocol (MCP) server seamlessly connecting AI Agents and AI coding tools with Zilliz Cloud https://zilliz.com/

一句话选型结论

按资料字段整理

Zilliz MCP Server 属于数据服务类 MCP 服务器,stdio · 本地进程,Token,可直接接入 Claude、Cursor、Windsurf,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类数据服务 传输方式stdio · 本地进程 鉴权Token 部署方式本地部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

数据服务
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

mcpragzillizmilvusdatabase

安装与配置

stdio · 本地进程 Token
客户端在本机拉起子进程,数据不出本机
ZILLIZ_CLOUD_TOKEN={ZILLIZ_CLOUD_TOKEN} uvx zilliz-mcp-server {"ZILLIZ_CLOUD_TOKEN":"your-token-here"}

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 Node.js 等运行时,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
数据库查询

用自然语言查库、取数,不用手写 SQL。

数据同步与清洗

把多来源数据拉到一处做整理。

报表取数

按口径快速拉出统计结果。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude、Cursor、Windsurf,换客户端不用重新研究配置。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 使用前要准备好Token对应的凭据,并按最小权限配置。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在数据服务场景里让模型直接调用外部能力的人。
  • 已经在用 Claude、Cursor 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Zilliz MCP Server当前 未标注 stdio · 本地进程 本地部署 有 2026-05-18
Memory 9d25f64e 9 未标注 未标注 无 2026-05-17
Getzep MCP 9 stdio · 本地进程 本地部署 有 2026-05-17
anything-llm 未标注 stdio · 本地进程 本地部署 有 2026-04-11
Prisma MCP Server 未标注 stdio · 本地进程 本地部署 有 2026-05-17
Astr Bot 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

数据服务
字段来自项目仓库与公开索引,中文说明由本站标注
提供方zilliztech上游标注的维护方 zilliztech
分类数据服务本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式Token远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · 未标注本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验2026-05-18上游最后核验时间

详细介绍

7.4 KB
来自项目仓库与上游目录的原文,未做删改

Zilliz-MCP-Server

1. Zilliz-MCP-Server Overview

Model Context Protocol (MCP) is a standardized framework that enables AI applications to securely connect to external data sources and tools in real-time. It acts as a universal interface between AI models and various systems, allowing AI assistants to access current information and perform actions beyond text generation.

Zilliz MCP Server enables AI agents to seamlessly interact with Milvus, a popular open-source vector database, and Zilliz Cloud, the fully managed version of Milvus. Through this integration, your AI assistants can create collections, insert vector data, and perform semantic searches directly within their conversations—no manual database management required. Zilliz MCP Server seamlessly integrates with popular AI-powered coding tools like Cursor, Claude, Windsurf, and other MCP-compatible editors, enabling developers to build vector search capabilities directly within their development workflow.

2. Demos

Demo 1: Create a free Milvus cluster with a simple natural language prompt

Instead of navigating through web interfaces or complex setup processes, you can get a free, fully functional vector database cluster up and running it directly from Claude, Cursor or any other MCP-compatible AI coding assistants with the Zilliz MCP Server.

The Zilliz MCP server automatically:
- Provisions a free Milvus cluster using Zilliz Cloud
- Handles authentication and configuration
- Returns connection details for immediate use

No need to leave your AI chat interface or manually set up infrastructure. Just ask in natural language and get a working vector database within seconds. This works in any MCP-enabled environment, including Claude's web interface as shown above.

Demo 2: Monitor cluster performance without leaving your chat

Once you have a cluster running, you can inspect its status and generate visualizations directly through natural language:

The Zilliz MCP server:
- Retrieves real-time cluster metrics and collection details
- Fetches performance data (CPU computation, capacity usage)
- Generates visualizations on demand
- All without writing queries or accessing monitoring dashboards

This demonstrates how you can monitor and analyze your vector database infrastructure conversationally, making cluster management as simple as asking questions in plain English.

Demo 3: Perform semantic search directly from your chat

Search your vector collections using natural language without writing any code:

The Zilliz MCP server:
- Performs semantic search across your specified collection
- Returns relevant results with similarity scores
- All through conversational interface - no SDK calls or query syntax required

This shows how vector search becomes as simple as describing what you're looking for. Perfect for testing search queries, exploring your data, or building search functionality without leaving your development environment.

3. Requirements

  • Python: 3.10 or higher.
  • uv: A fast Python package installer and resolver. If you don't have it, run brew install uv on OSX, or just run curl -LsSf https://astral.sh/uv/install.sh | sh .
  • A Zilliz Cloud account: sign up for free if you haven't tried Zilliz Cloud before. If you already have one, sign in here.
  • Zilliz Cloud API Key: You'll need an API key to interact with Zilliz Cloud. You can get one by following the instructions here: Manage API Keys.

4. Usage

You can start the server in two ways:

If you are completely new to MCP and Zilliz, we recommend following our step-by-step guide in Step By Step User Guide which will walk you through the complete setup process.

4.1. Standard I/O (StdIO)

This method is useful when the agent and the MCP server are running on the same machine and you want the agent to manage the server's lifecycle directly. The agent communicates with the server over its standard input and output streams.

Configure your agent's MCP JSON file like this:```json
{
"mcpServers": {
"zilliz-mcp-server": {
"command": "uvx",
"args": ["zilliz-mcp-server"],
"env": {
"ZILLIZ_CLOUD_TOKEN": "your-token-here"
}
}
}
}

*Note: Make sure to replace `/path/to/your/zilliz-mcp-server` with the actual absolute path to the project directory.*

### 4.2. Streamable HTTP

This method runs the server as a standalone HTTP service. This is useful for development and for agents that can communicate over HTTP.

First, clone or download the project repository.```bash
git clone https://github.com/zilliztech/zilliz-mcp-server.git
cd zilliz-mcp-server

Next, create a .env file from the example and fill in your Zilliz Cloud API key.```bash
cp example.env .env


Now, open `.env` and add your API key:```
ZILLIZ_API_KEY="your_api_key_here"

It 'll start the MCP server as a standalone HTTP service```bash
uv run src/zilliz_mcp_server/server.py --transport streamable-http


After starting the server, you can configure your MCP client to connect to it. If the server is running correctly, the available tools will appear in your client's tool list (e.g., in Cursor or Claude).

You can then configure your agent or MCP client to connect to it using a configuration like this:```json
{
  "mcpServers": {
    "zilliz-mcp-server": {
      "url": "http://localhost:8000/mcp",
      "transport": "streamable-http",
      "description": "Zilliz Cloud and Milvus MCP Server"
    }
  }
}

5. Available Tools

The server exposes two categories of tools for your AI agents.

Zilliz Control Plane Tools

These tools are for managing your Zilliz Cloud resources.

Tool Name Description
list_projects List all projects in your Zilliz Cloud account.
list_clusters List all clusters within your projects.
create_free_cluster Create a new, free-tier Milvus cluster.
describe_cluster Get detailed information about a specific cluster.
suspend_cluster Suspend a running cluster to save costs.
resume_cluster Resume a suspended cluster.
query_cluster_metrics Query various performance metrics for a cluster.

Milvus Data Plane Tools

These tools are for interacting with the data inside a Milvus cluster.

Tool Name Description
list_databases List all databases within a specific cluster.
list_collections List all collections within a database.
create_collection Create a new collection with a specified schema.
describe_collection Get detailed information about a collection, including its schema.
insert_entities Insert entities (data records with vectors) into a collection.
delete_entities Delete entities from a collection based on IDs or a filter expression.
search Perform a vector similarity search on a collection.
query Query entities based on a scalar filter expression.
hybrid_search Perform a hybrid search combining vector similarity and scalar filters.

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
Zilliz MCP Server 怎么安装?

在客户端按下面的安装命令接入:ZILLIZ_CLOUD_TOKEN={ZILLIZ_CLOUD_TOKEN} uvx zilliz-mcp-server {"ZILLIZ_CLOUD_TOKEN":"your-token-here"}。装完在客户端配置里引用同一个包或端点。

Zilliz MCP Server 怎么在 Claude、Cursor、Windsurf 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Zilliz MCP Server 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Zilliz MCP Server 需要 API Key 吗?

资料标注需要Token认证,使用前准备好对应凭据,并按最小权限配置。

Zilliz MCP Server 支持哪些客户端?

资料标注支持 Claude、Cursor、Windsurf;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Zilliz MCP Server 能提供哪些能力?

资料显示能力标签包括 数据分析、RAG、Claude。具体工具清单以官方文档为准。

Zilliz MCP Server 需要什么运行环境?

属于本地进程,需要在本机装好运行环境。

Zilliz MCP Server 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-18。