kdash

A terminal-based Kubernetes dashboard with AI integration. Navigate clusters, inspect resources, tail logs, and ask AI questions — all without leaving the terminal.

⎈ kdash ─ kubernetes dashboard

╭─────────────────────────────────────────────────────────╮
│ ⎈ minikube │ staging │ Deployments          3 items     │
╰─────────────────────────────────────────────────────────╯
┌─────────────────────────────────────────────────────────┐
│ ◆ 1:Deploy  2:STS  3:DS  4:Pods  5:Job ...             │
│                                                         │
│  NAME          READY  UP-TO-DATE  AVAILABLE  AGE        │
│ ❯ nginx        3/3    3           3          5d12h      │
│   redis        1/1    1           1          5d12h      │
│   api-gateway  2/2    2           2          3d4h       │
└─────────────────────────────────────────────────────────┘
j/k nav  ↵ open  1-0 tab  [] cycle  n ns  l logs  a ai

Why

Web-based Kubernetes dashboards require complex authentication flows and context switching. As a developer, you already have kubectl access in your terminal. kdash puts a navigable, keyboard-driven resource browser right where you work.

Features

• 14 resource types — Deployments, StatefulSets, DaemonSets, Pods, Jobs, CronJobs, Services, Ingresses, ConfigMaps, Secrets, PVCs, PVs, Events, Nodes
• Keyboard-driven — vim-style navigation (j/k, g/G), number keys for tabs, drill-down with Enter/Esc
• Real-time log tailing — press l on any pod to stream logs with pause, scroll, and previous container support
• AI analysis — press a to ask Claude about any resource (auto-includes kubectl describe context)
• Copyable commands — press y to get context-aware kubectl commands copied to clipboard
• Read-only — no mutations from the UI. Dangerous operations are surfaced as copyable commands
• Namespace switching — press n to switch namespaces instantly
• Drill-down with sub-tabs — enter a Deployment to see its Overview, Pods, Events, and YAML
• Filter — press / to filter any resource list

Install

npm install -g kdash

Or run directly:

npx kdash

Prerequisites

• Node.js 20+
• kubectl configured with cluster access
• claude CLI (optional, for AI features) — install

Usage

kdash

Select a Kubernetes context, then navigate:

Keybindings

Key	Action
j/k ↑/↓	Navigate rows
Enter	Drill into resource
Esc	Go back
1-9 0	Switch to tab 1-10
[ ]	Cycle tabs (prev/next)
g / G	Jump to top / bottom
/	Filter current list
n	Switch namespace
l	Tail pod logs
a	Ask AI about resource
y	Copy kubectl command
f	Fullscreen panel
p / P	Pause logs / previous container
Tab	Switch pane focus
?	Show help
q	Quit

Resource Tabs

#	Tab	Category
1	Deployments	Workloads
2	StatefulSets	Workloads
3	DaemonSets	Workloads
4	Pods	Workloads
5	Jobs	Batch
6	CronJobs	Batch
7	Services	Networking
8	Ingresses	Networking
9	ConfigMaps	Config
0	Secrets	Config
]	PVCs, PVs	Storage
]	Events, Nodes	Cluster

AI Integration

Press a on any resource to ask Claude a question. kdash automatically runs kubectl describe and passes the output as context. Follow-up questions maintain conversation history (last 5 exchanges).

Requires the Claude CLI installed and authenticated.

Log Viewer

Press l on a pod to tail its logs in real-time. The log panel opens at the bottom:

• j/k — scroll through logs
• g/G — jump to top/bottom
• f — fullscreen
• p — pause/resume tailing
• P — show previous container logs (for crashed pods)
• Esc — close

Development

git clone https://github.com/damahua/kdash.git
cd kdash
npm install

# Set up a local test cluster
bash scripts/setup-test-cluster.sh

# Run in dev mode
npm run dev

# Run tests
npm test

# Build
npm run build

Tech Stack

• TypeScript + Ink (React for terminals)
• @kubernetes/client-node — official Kubernetes JS client
• claude CLI — AI analysis via claude -p

License

MIT