diff --git a/content/en/news/ks-mcp-server-2025.md b/content/en/news/ks-mcp-server-2025.md new file mode 100644 index 000000000..23fcac669 --- /dev/null +++ b/content/en/news/ks-mcp-server-2025.md @@ -0,0 +1,149 @@ +--- +title: 'KubeSphere MCP Server: Enhancing AI Integration with KubeSphere' +tag: 'Product News' +keyword: 'open source, KubeSphere, K8s, MCP' +description: 'KubeSphere MCP Server enables seamless integration of AI assistants with KubeSphere, allowing users to manage cloud-native resources through natural language interactions.' +createTime: '2025-5-6' +author: 'KubeSphere' +image: 'https://pek3b.qingstor.com/kubesphere-community/images/KSMCPSERVER2025.png' +--- + + +As cloud-native and AI technologies rapidly converge, the open-source community has been searching for efficient ways to seamlessly integrate AI capabilities into cloud platforms. We are excited to announce that KubeSphere now supports the Model Context Protocol (MCP). With the new KubeSphere MCP Server, users can interact with the KubeSphere API via AI assistants, enabling a smarter cloud management experience. + +## What is KubeSphere MCP Server? + +KubeSphere MCP Server is a service implemented based on the Model Context Protocol (MCP) specification. It enables seamless integration with the KubeSphere API, allowing AI assistants to access and operate on KubeSphere resources. The server consists of four main functional modules: + +- **Workspace Management** - Create, query, and manage KubeSphere workspaces +- **Cluster Management** - Monitor and manage Kubernetes cluster resources +- **User and Role Management** - Manage user permissions and role assignments +- **Extension Center** - Extend the functionality of KubeSphere + +With these modules, users can perform actions originally done via the KubeSphere console or command line by simply conversing with an AI assistant, significantly improving operational efficiency. + +## Why Use KubeSphere MCP Server? + +- **Natural Language Interaction**：With MCP-supported AI assistants such as Claude or Cursor, users can interact with KubeSphere using natural language +- **Lower Learning Curve**：Easily manage resources even without familiarity with KubeSphere APIs or CLI +- **Efficient Troubleshooting**：AI assistants can analyze cluster health and suggest solutions +- **Workflow Automation**：Automate multi-step operations via conversational interfaces + +## Quick Start Guide + +### Prerequisites + +- A deployed KubeSphere cluster +- Access credentials including address, username, and password + +### Setup Steps + +#### Step 1: Generate KubeSphere Configuration File + +The configuration file is similar to a kubeconfig and includes HTTP connection details. The default context name is `kubesphere`, which can be modified using the environment variable `KUBESPHERE_CONTEXT`. + +```yaml +apiVersion: v1 +clusters: +- cluster: + certificate-authority-data: + server: + name: kubesphere +contexts: +- context: + cluster: kubesphere + user: admin + name: kubesphere +current-context: kubesphere +kind: Config +preferences: {} +users: +- name: admin + user: + username: + password: +``` + +**Note:** + +- ``: Optional. Provide the base64-encoded CA cert if using HTTPS +- ``: Required. Must be an HTTPS address (can be overwritten using --ks-apiserver for HTTP) +- `` and ``: Required for authentication + +#### Step 2: Obtain the `ks-mcp-server` Binary + +You can obtain the binary by: + +- Building from source: `go build -o ks-mcp-server cmd/main.go` +- Downloading from [GitHub Releases](https://github.com/kubesphere/ks-mcp-server/releases) + +Ensure the binary is added to your system's `$PATH`. + +#### Step 3: Configure MCP Server in AI Clients + +**Claude Desktop Configuration** + + +```json +{ + "mcpServers": { + "KubeSphere": { + "args": [ + "stdio", + "--ksconfig", "", + "--ks-apiserver", "" + ], + "command": "ks-mcp-server" + } + } +} +``` + + +**Cursor Configuration** + + +```json +{ + "mcpServers": { + "KubeSphere": { + "command": "ks-mcp-server", + "args": [ + "stdio", + "--ksconfig","", + "--ks-apiserver","" + ] + } + } +} +``` +``: Required. Absolute path to the config file created in Step 1 +``: Optional unless using HTTP; specify `ks-console` or `ks-apiserver` address like `http://172.10.0.1:30880` + +### Usage Examples + +Once setup is complete, you can interact with KubeSphere through AI assistants like Claude Desktop or Cursor. Example interactions include: + +- "List all workspaces in KubeSphere" +![](https://pek3b.qingstor.com/kubesphere-community/images/list%20all%20workspaces%20mcp.png) + +- "Show my cluster in KubeSphere" +![](https://pek3b.qingstor.com/kubesphere-community/images/cluster%20in%20kubesphere%20mcp.png) + +The AI assistant will retrieve the information via KubeSphere MCP Server and present it in human-readable format. + +## Community Participation + +KubeSphere MCP Server is open source, and contributions or feedback from the community are welcome. You can reach out via: + +- GitHub：[https://github.com/kubesphere/ks-mcp-server](https://github.com/kubesphere/ks-mcp-server) +- Slack Channel: [https://kubesphere.slack.com](https://kubesphere.slack.com) + +## Conclusion + +KubeSphere MCP Server represents a forward-looking fusion of cloud-native management and AI interaction. By integrating powerful language model capabilities with the rich functionality of KubeSphere, we aim to deliver a more intelligent and efficient cloud management experience. Whether you are a seasoned DevOps engineer or new to cloud-native technologies, the MCP Server can help you manage and operate the KubeSphere platform with ease. + +**Start using KubeSphere MCP Server now and experience the AI-powered future of cloud-native operations!** + + + diff --git a/content/zh/news/ks-mcp-serverzh.md b/content/zh/news/ks-mcp-serverzh.md new file mode 100644 index 000000000..2e1c4543d --- /dev/null +++ b/content/zh/news/ks-mcp-serverzh.md @@ -0,0 +1,150 @@ +--- +title: 'KubeSphere MCP Server: 增强AI与KubeSphere的集成能力' +tag: '产品动态' +keyword: '社区, 开源, MCP, KubeSphere' +description: 'KubeSphere MCP Server 实现了 AI 助手与 KubeSphere 的无缝集成，使用户能够通过自然语言交互管理云原生资源。' +createTime: '2025-5-6' +author: 'KubeSphere' +image: 'https://pek3b.qingstor.com/kubesphere-community/images/KSMCPSERVER2025.png' +--- + + +在云原生和AI技术快速融合的今天，开源社区一直在寻找将 AI 能力无缝集成到云平台的高效方式。我们很高兴地宣布，KubeSphere 现已支持Model Context Protocol (MCP)，通过全新的 KubeSphere MCP Server，用户可以使用 AI 助手直接与 KubeSphere API 交互，实现更智能化的云平台管理体验。 + + +## 什么是KubeSphere MCP Server？ + +KubeSphere MCP Server是基于 **[Model Context Protocol (MCP)]** 规范实现的服务，它提供了与KubeSphere API的无缝集成，使AI助手能够直接获取和操作KubeSphere资源。该服务主要分为四个功能模块： + +- **工作空间管理** - 创建、查询和管理 KubeSphere 工作空间 +- **集群管理** - 监控和管理 K8s 集群资源 +- **用户和角色** - 管理用户权限和角色分配 +- **扩展中心** - 扩展 KubeSphere 的功能 + +通过这些模块，用户可以简单地通过与 AI 助手的对话来完成原本需要在 KubeSphere 控制台或命令行中进行的操作，大大提高管理效率。 + + +## 为什么使用KubeSphere MCP Server？ + +- **自然语言交互**：通过 Claude、Cursor 等支持 MCP 的 AI 助手，用户可以使用自然语言与 KubeSphere 平台交互 +- **降低学习门槛**：即使不熟悉 KubeSphere API或命令行操作，也能轻松管理资源 +- **高效问题诊断**：AI 助手可以帮助分析集群状态，提供问题解决建议 +- **流程自动化**：通过对话式界面自动执行多步骤操作 + +## 快速上手指南 + +### 前提条件 + +已部署的 KubeSphere 集群（包含访问地址、用户名和密码） + + +### 配置步骤 + +#### 1. 生成KubeSphere配置文件 + +配置文件格式类似于 kubeconfig，包含 HTTP 连接信息。默认的 KubeSphere context名称为`kubesphere`，可以通过环境变量`KUBESPHERE_CONTEXT`修改。 + +```yaml +apiVersion: v1 +clusters: +- cluster: + certificate-authority-data: + server: + name: kubesphere +contexts: +- context: + cluster: kubesphere + user: admin + name: kubesphere +current-context: kubesphere +kind: Config +preferences: {} +users: +- name: admin + user: + username: + password: +``` +``: 非必填，当KubeSphere使用HTTPS访问时，填写base64编码格式的CA证书 +``: 必填，必须填写HTTPS的地址。（如果使用HTTP，此处填写任意HTTPS的地址，然后通过参数`--ks-apiserver http://xxx`进行修改） +``: 必填，KubeSphere集群的用户 +``: 必填，KubeSphere集群的用户的密码 + +#### 2. 获取ks-mcp-server二进制文件 + +您可以通过以下方式获取 ks-mcp-server 二进制文件： + +- 从源码构建：`go build -o ks-mcp-server cmd/main.go` +- 从[GitHub Releases](https://github.com/kubesphere/ks-mcp-server/releases)下载 + +获取后，请确保将其添加到系统的`$PATH`环境变量中。 + +#### 3. 在AI客户端中配置MCP服务器 + +**Claude Desktop配置** + +按照 [Claude Desktop文档](https://modelcontextprotocol.io/quickstart/user) 修改MCP配置： + +```json +{ + "mcpServers": { + "KubeSphere": { + "args": [ + "stdio", + "--ksconfig", "", + "--ks-apiserver", "" + ], + "command": "ks-mcp-server" + } + } +} +``` +``: 必填，通过步骤 **1-生成 KubeSphere 配置文件** 的绝对路径。 +``: 非必填，但HTTP访问时必填，KubeSphere集群的访问地址，支持ks-console或ks-apiserver的服务地址，例如：`http://172.10.0.1:30880` + +**Cursor配置** + +按照 [Cursor文档](https://docs.cursor.com/context/model-context-protocol) 修改MCP配置： + +```json +{ + "mcpServers": { + "KubeSphere": { + "command": "ks-mcp-server", + "args": [ + "stdio", + "--ksconfig","", + "--ks-apiserver","" + ] + } + } +} +``` +``: 必填，通过步骤 **1-生成 KubeSphere 配置文件** 的绝对路径。 +``: 非必填，但HTTP访问时必填，KubeSphere集群的访问地址，支持ks-console或ks-apiserver的服务地址，例如：`http://172.10.0.1:30880` + +### 使用示例 + +配置完成后，您可以通过 Claude Desktop 或 Cursor 等AI助手与 KubeSphere 交互。示例对话： + +- "列出 KubeSphere 中所有的工作空间" +![](https://pek3b.qingstor.com/kubesphere-community/images/ksmcpclusterchatu1.png) + +- "显示我的集群状态" +![](https://pek3b.qingstor.com/kubesphere-community/images/ksmcpclusterchatu2.png) + +AI助手将通过 KubeSphere MCP Server 获取相关信息并以人类可读的方式呈现。 + +## 社区参与 + +KubeSphere MCP Server 是一个开源项目，我们欢迎社区成员的贡献和反馈。如果您有任何问题或建议，请通过以下方式联系我们： + +- GitHub仓库：[https://github.com/kubesphere/ks-mcp-server](https://github.com/kubesphere/ks-mcp-server) +- KubeSphere论坛：[https://kubesphere.com.cn/forum/](https://kubesphere.com.cn/forum/) +- Slack频道：[https://kubesphere.slack.com](https://kubesphere.slack.com) + +## 结语 + +KubeSphere MCP Server 代表了云原生管理与AI交互的未来方向。通过将强大的语言模型能力与 KubeSphere 的丰富功能相结合，我们致力于为用户提供更智能、更高效的云平台管理体验。无论您是资深 DevOps 工程师还是刚接触云原生的新手，KubeSphere MCP Server 都能帮助您更轻松地管理和使用KubeSphere平台。 + +现在就开始尝试 KubeSphere MCP Server，体验AI赋能的云原生管理新方式！ \ No newline at end of file