Repolode
Submit
weibaohui

weibaohui/k8m

A lightweight, single-binary Kubernetes dashboard with built-in AI assistance, MCP tool integration, and multi-cluster management.

Open-source alternative to:LensRancher

View on GitHubWebsite
Stars
892
Forks
164
Open issues
25
Last commit
4 hours ago

What is it?

k8m is a Mini Kubernetes dashboard built on AMIS, using the kom library as its Kubernetes API client. It is distributed as a single executable with the frontend embedded, and it runs on Linux, macOS and Windows across x86 and ARM. It works with standard Kubernetes as well as EKS, k3s, kind and k0s clusters, and can register and monitor several clusters at once with heartbeat checks and automatic reconnection.

The AI features connect to a user-supplied model API key, including self-hosted options such as Ollama; no model is bundled. They cover explaining selected text, translating YAML fields, interpreting resource descriptions, analyzing pod logs and recommending commands. k8m also exposes its Kubernetes operations as MCP tools, and each MCP call runs with the permissions of the user who invoked it.

Beyond AI, k8m provides pod file browsing and editing, log viewing and shell access, per-cluster role authorization with namespace allow and deny lists, scheduled cluster inspections with alerts to DingTalk, WeCom, Feishu or webhooks, Kubernetes event forwarding, CRD discovery, and a Helm chart manager.

Key features

  • 01Single-binary deployment with the web frontend embedded
  • 02Multi-cluster registration with heartbeat checks and auto-reconnect
  • 03AI-assisted YAML translation, describe interpretation and log analysis using your own model keys
  • 04MCP server exposing Kubernetes operations as tools, with calls run under the invoking user's permissions
  • 05Per-cluster read-only, exec and admin authorization with namespace filters
  • 06Scheduled cluster inspections and Kubernetes event forwarding to chat webhooks

Use cases

  • Operating several Kubernetes clusters from one lightweight web console
  • Letting an LLM agent inspect or operate a cluster through MCP with controlled permissions
  • Running a self-hosted cluster dashboard with Ollama instead of a cloud AI service
  • Sending scheduled health reports or cluster events to a team chat group

Quick start

  1. 1

    Download the latest release from the GitHub releases page

  2. 2

    Start k8m with the binary

    ./k8m
  3. 3

    Open http://127.0.0.1:3618 and log in with the default username and password k8m / k8m

Things to know

  • AI features require your own model API key, since k8m does not ship a built-in model.
  • The default login credentials (k8m / k8m) should be changed before going live.

In collections

MCP serversRun LLMs locallyDashboards & business intelligence toolsLogin, SSO & identity servers

Topics

#kubernetes#dashboard#mcp#llm#multi-cluster#devops-tools#ollama#helm#2fa#ai

FAQ

What is k8m?
A lightweight, single-binary Kubernetes dashboard with built-in AI assistance, MCP tool integration, and multi-cluster management. k8m is a Mini Kubernetes dashboard built on AMIS, using the kom library as its Kubernetes API client. It is distributed as a single executable with the frontend embedded, and it runs on Linux, macOS and Windows across x86 and ARM. It works with standard Kubernetes as well as EKS, k3s, kind and k0s clusters, and can register and monitor several clusters at once with heartbeat checks and automatic reconnection.
Is k8m free to use? What license does it have?
k8m is open source under the MIT license. Check the license terms before using it in commercial products.
Is k8m still maintained?
Actively maintained. The last commit to k8m was on Nov 7, 2026.
What can I use k8m for?
Operating several Kubernetes clusters from one lightweight web console Letting an LLM agent inspect or operate a cluster through MCP with controlled permissions Running a self-hosted cluster dashboard with Ollama instead of a cloud AI service Sending scheduled health reports or cluster events to a team chat group
How do I install k8m?
Download the latest release from the GitHub releases page Start k8m with the binary: ./k8m Open http://127.0.0.1:3618 and log in with the default username and password k8m / k8m
Where can I get k8m?
The source code and installation instructions are on GitHub: https://github.com/weibaohui/k8m

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects