AI KUBERNETES 콘솔

KudashAI: 자주적
에이전트 운영

개인정보 보호를 최우선으로 하는 엔터프라이즈 클러스터 운영을 위한 Plan-Execute-Verify 에이전트 및 로컬 LLM 지원을 갖춘 AI 기반 멀티 클러스터 Kubernetes 콘솔.

$ kudash-agent plan "scale deployment/web to 5 replicas"

Analyzing cluster state...

Proposed Plan:

1. Patch Deployment: web

2. replicas: 3 -> 5

3. Check HPA constraints... OK

Waiting for operator authorization...

STATUS: IDLE_WAIT_EXEC APPROVE

Plan-Execute-Verify

KudashAI operates as a virtual SRE. Every cluster mutation follows a controlled three-phase workflow that keeps the operator in command.

Describe a task in plain text. The AI drafts a clear execution plan listing every Kubernetes mutation it intends to perform.

psychology NATURAL LANGUAGE ANALYSIS

The plan is presented in the interface. No mutations occur until the operator clicks the confirmation button, maintaining full control.

verified_user HUMAN-IN-THE-LOOP

Once authorized, the AI executes API calls, reads updated pod statuses, and reports the outcome with full audit detail.

visibility FULL AUDIT LOGGING

High-performance infrastructure meets intelligent automation. KudashAI is built for the stringent requirements of global enterprise.

shield_lock

Fully compatible with local LLM runtimes including Ollama and LMStudio. Run AI models on-premise without sending cluster data to external APIs.

OLLAMA LMSTUDIO

cloud_sync

Support for OpenAI, Anthropic, Gemini, DeepSeek, AWS Bedrock, and OpenRouter when cloud-hosted inference is preferred.

desktop_windows

Native desktop application compiled via Wails (Go + Webview) with built-in auto-updaters for macOS, Linux, and Windows.

bolt

Securely connect local ports directly to in-cluster Kubernetes services with a single action from the console interface.

terminal

Monaco YAML editor and xterm.js-based WebSocket terminals to execute container commands directly from the dashboard.

code

Clean Radix UI and Tailwind CSS dashboard paired with a performant Go backend API server.

KudashAI brings high-density data visualization and agentic automation together. Read the documentation or explore the underlying host OS with Vapor.