Ks Mcp Server

Ks Mcp Server

What is Ks Mcp Server about?

It enables AI agents and other tools to retrieve and manipulate KubeSphere resources such as workspaces, clusters, users, roles, and extensions via a standardized Model Context Protocol interface.

How to use Ks Mcp Server?

1. Prepare a ksconfig file – a kubeconfig‑like YAML containing the KubeSphere endpoint, optional CA data, and user credentials.
2. Build or download the binary – go build -o ks-mcp-server cmd/main.go or fetch a released binary.
3. Configure the MCP server for your AI client (e.g., Claude Desktop, Cursor) by adding a JSON block:

{
  "mcpServers": {
    "KubeSphere": {
      "args": [
        "stdio",
        "--ksconfig", "<absolute path to ksconfig>",
        "--ks-apiserver", "<KubeSphere address>"
      ],
      "command": "ks-mcp-server"
    }
  }
}

4. Launch the AI client; it will communicate with the server through standard I/O.

Key features of Ks Mcp Server

• Workspace Management – list, create, update, and delete KubeSphere workspaces.
• Cluster Management – query clusters, retrieve status, and manage cluster resources.
• User & Role Operations – fetch user details, roles, and permission mappings.
• Extensions Center – access custom extensions and plugins registered in KubeSphere.
• MCP Compatibility – works with any tool that supports the Model Context Protocol, enabling seamless AI‑agent integration.

Use cases of Ks Mcp Server

• AI‑assisted DevOps – let conversational agents query cluster health or create namespaces on‑the‑fly.
• Automated Documentation – generate docs about current workspaces and resources by prompting an AI model.
• Self‑service Portals – embed an AI chat widget that can provision resources without manual UI navigation.
• Monitoring & Alerting Integration – feed real‑time cluster data to AI‑driven anomaly detection services.

FAQ from the Ks Mcp Server

Q: Do I need a KubeSphere cluster to run the server? A: Yes, the server requires a reachable KubeSphere cluster and valid credentials.

Q: Is HTTPS mandatory? A: The server field in ksconfig must be an HTTPS URL. For HTTP access, provide --ks-apiserver http://... at runtime.

Q: Can I run the server without building from source? A: Absolutely. Pre‑compiled binaries are available on the GitHub releases page.

Q: Which AI clients are supported? A: Any client that implements the Model Context Protocol, such as Claude Desktop and Cursor, can be configured as shown above.

KubeSphere MCP Server

The KubeSphere MCP Server is a Model Context Protocol(MCP) server that provides integration with KubeSphere APIs, enabling to get resources from KubeSphere. Divided into four tools modules: Workspace Management, Cluster Management, User and Roles, Extensions Center.

Prerequisites

You must have a KubeSphere cluster. contains: Access Address, Username, Password.

Installation

Generate KSConfig

The format is similar to kubeconfig. This configuration contains HTTP connector information. The default context for KubeSphere is kubesphere, which can be modified via the environment variable KUBESPHERE_CONTEXT.

apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: <CA file>
    server: <Server Address>
  name: kubesphere
contexts:
- context:
    cluster: kubesphere
    user: admin
  name: kubesphere
current-context: kubesphere
kind: Config
preferences: {}
users:
- name: admin
  user:
    username: <KubeSphere Username>
    password: <KubeSphere Password>

<CA file>: Optional. Fill in the CA certificate in base64-encoded format when KubeSphere is accessed via HTTPS.
<Server Address>: Required Must be an HTTPS address. (If using HTTP, enter any HTTPS address here, then modify via the parameter --ks-apiserver http://xxx)
<KubeSphere Username>: Required The user for the KubeSphere cluster.
<KubeSphere Password>: Required The password for the KubeSphere cluster user.

Get ks-mcp-server binary

you can run command go build -o ks-mcp-server cmd/main.go or download from (github releases)[https://github.com/kubesphere/ks-mcp-server/releases] and then move it to $PATH.

Configuration MCP Server in AI Agent

Claude Desktop

1. According to Claude Desktop should change the MCP Configuration. like:

{
  "mcpServers": {
    "KubeSphere": {
      "args": [
        "stdio",
        "--ksconfig", "<ksconfig file absolute path>",
        "--ks-apiserver", "<KubeSphere Address>"
      ],
      "command": "ks-mcp-server"
    }
  }
}

<ksconfig file absolute path>: Required The absolute path of the ksconfig file.
<KubeSphere Address>: Optional (but required for HTTP access) The access address of the KubeSphere cluster, supporting either the ks-console or ks-apiserver service address (e.g., http://172.10.0.1:30880).

2. chat with mcp server

Cursor

1. According to Curosr should change the MCP Configuration. like:

{
  "mcpServers": {
    "KubeSphere": {
      "args": [
        "stdio",
        "--ksconfig", "<ksconfig file absolute path>",
        "--ks-apiserver", "<KubeSphere Address>"
      ],
      "command": "ks-mcp-server"
    }
  }
}

<ksconfig file absolute path>: Required The absolute path of the ksconfig file.
<KubeSphere Address>: Optional (but required for HTTP access) The access address of the KubeSphere cluster, supporting either the ks-console or ks-apiserver service address (e.g., http://172.10.0.1:30880).

2. chat with mcp server