weibaohui/k8m

Free

提供 MCP 多集群 Kubernetes 管理和操作，带管理界面、日志记录和近 50 个内置工具 (支持标准/CRD)。

FreeFree tier

Type

Open Source

Visit Website GitHub (40 stars)

About weibaohui/k8m

K8m is an AI-driven Mini Kubernetes AI Dashboard designed to simplify multi-cluster management through a lightweight, single-binary deployment. Built with Go and the AMIS frontend framework, it integrates a built-in Qwen2.5-Coder-7B model (also supporting DeepSeek-R1-Distill-Qwen-7B) and allows connection to private LLMs including Ollama. Key capabilities include MCP (Model Context Protocol) integration with 49 built-in tools for cluster operations, permission-aware MCP execution, multi-cluster auto-discovery, RBAC with cluster-level roles, Pod file management, real-time log streaming with search, cluster inspection with custom Lua rules, Event forwarding with AI summarization, CRD auto-discovery, Helm marketplace, and API key generation with Swagger documentation. It supports standard Kubernetes, AWS EKS, k3s, kind, k0s, and runs on Linux, macOS, Windows, and multiple architectures. The project is fully open source under the Apache 2.0 license.

Key Features

Miniaturized single-file deployment (single binary)

Built-in Qwen2.5-Coder-7B LLM with support for DeepSeek-R1-Distill-Qwen-7B and private Ollama models

MCP integration with 49 built-in Kubernetes management tools

MCP permission system linking cluster RBAC to AI model calls

Multi-cluster auto-discovery, heartbeat detection, and automatic reconnection

Role-based access control per cluster (read-only, exec, admin) with namespace allow/block lists

Pod file management (browse, edit, upload, download, delete) via console file tree

Real-time pod log viewing with Ctrl+F search similar to grep -A -B

Cluster inspection with scheduled checks, custom Lua rules, and alerts to DingTalk, WeChat, Feishu, or webhooks

Kubernetes Event forwarding with filtering by cluster, keyword, namespace, and AI summarization

Pros & Cons

Pros

Completely open source and free to use with no restrictions
AI-driven features reduce manual intervention and speed up cluster troubleshooting
Single binary deployment eliminates complex setup and dependencies
Comprehensive RBAC and MCP permission system ensures secure AI-driven cluster access
Supports a wide variety of Kubernetes distributions and cloud providers

Cons

Initial configuration requires kubeconfig and LLM endpoint setup
Built-in LLM (7B parameters) may be less capable than larger models for complex reasoning
Documentation is primarily in Chinese, which may present a barrier for non-Chinese users
User interface based on AMIS may have a learning curve compared to native dashboards

Best For

Simplified multi-cluster Kubernetes management for DevOps and SRE teamsAI-assisted troubleshooting with log analysis, YAML translation, and command recommendationsAutomated cluster operations and inspection via MCP-driven AI toolsEducational environment for learning Kubernetes with built-in AI guidance and gloss explanations

FAQ

What is k8m?

K8m is an AI-driven Mini Kubernetes Dashboard that integrates multi-cluster management, MCP tools, and built-in large language models to simplify cluster operations.

Is k8m free to use?

Yes, k8m is entirely open source under the Apache 2.0 license and free to use, including commercial use.

Can I use my own LLM instead of the built-in one?

Yes, k8m supports connecting to private LLMs, including those served by Ollama, as well as DeepSeek models.

What Kubernetes distributions are supported?

K8m supports standard Kubernetes, AWS EKS, k3s, kind, k0s, and other distributions that provide a kubeconfig file.

What is MCP integration?

MCP (Model Context Protocol) integration allows large language models to call 49 built-in Kubernetes management tools, enabling AI-driven cluster operations with permission control.