← Back to Discover

KubeStellar Console

Open-source dashboard for managing several Kubernetes clusters in one place, with optional AI help, guided install missions for CNCF projects and a local agent that bridges to your clusters.

Workflow automationAutonomous agents

What it does

KubeStellar Console is an open-source web dashboard, released under the Apache License 2.0, for working with multiple Kubernetes clusters together. A hosted demo with built-in sample data lets you explore the interface without a cluster or any install. To work with real clusters you self-host the console and run kc-agent, a small local daemon that connects the browser to your kubeconfig contexts and to AI providers. The agent can use CLI-based coding agents already installed on the machine, as well as self-hosted model servers such as Ollama, and it connects to MCP servers that query cluster state. AI is optional: without a provider configured, the dashboards and missions still work and AI features fall back to rule-based behaviour. The console provides guided install missions for more than 250 CNCF projects, with pre-flight checks, validation, troubleshooting and rollback. It also offers adaptive card suggestions, live event streaming from connected clusters, and a marketplace of community dashboards. An alpha runtime called Stellar adds persistent missions, operational memory and event-driven triggers.

Capabilities

  • Multi-cluster Kubernetes dashboard with live event streaming
  • Guided install missions for 250+ CNCF projects
  • kc-agent bridge to kubeconfig contexts, AI providers and MCP servers
  • Works with local or self-hosted models such as Ollama
  • Marketplace for community dashboards, card presets and themes
  • Stellar alpha runtime for persistent missions and triggers

When to use it

  • You manage more than one Kubernetes cluster and want a single dashboard
  • You want step-by-step guided installs of CNCF projects with validation and rollback
  • You want AI assistance for cluster operations from tools you already run locally
  • You need to run without external AI access, since AI features are optional