A keyboard-driven terminal UI for Kubernetes operations.

Current release: 1.1.24.

# Homebrew
brew install kxxil01/tap/kubectui

# From source
git clone https://github.com/kxxil01/Kubectui.git
cd Kubectui && cargo build --release
./target/release/kubectui

Requires a kubeconfig at ~/.kube/config (or KUBECONFIG env var).

1. Launch kubectui in a shell that already targets the cluster you want to inspect.
2. Press ? for in-app keybinding help and : for the action palette.
3. Use c to switch kube context, ~ to switch namespace, and / to filter the current view.
4. Open detail with Enter, then use the detail actions or the bottom workbench to inspect and mutate safely.

Persistent app preferences are written to ~/.kube/kubectui-config.json.

Optional command extensions load from:

$XDG_CONFIG_HOME/kubectui/extensions.yaml
# or, if XDG_CONFIG_HOME is unset:
~/.config/kubectui/extensions.yaml

Minimal example:

actions:
  - id: describe_pod
    title: Describe with kubectl
    resource_kinds: ["Pod"]
    aliases: ["describe pod"]
    mode: background
    command:
      program: kubectl
      args: ["describe", "pod", "$NAME", "-n", "$NAMESPACE"]
      timeout_secs: 120

timeout_secs applies to background and silent extension commands. Default is 120 seconds; accepted range is 1 to 3600 seconds. foreground commands run in your terminal until they exit.

Native AI actions load from ~/.kube/kubectui-config.json:

{
  "namespace": "all",
  "ai": {
    "providers": [
      { "provider": "codex_cli" },
      { "provider": "claude_cli" }
    ]
  }
}

Supported AI providers: open_ai, anthropic, claude_cli, and codex_cli. When multiple providers are configured, the action palette is the picker: choose entries like Explain Failure (Codex CLI) or Explain Failure (Claude CLI). Legacy ai: blocks in extensions.yaml are not supported.

Pod exec shells use container-local shells. Configure preferred candidates in ~/.kube/kubectui-config.json; login only affects shells that support login mode (bash, zsh, fish):

{
  "exec": {
    "shells": ["/usr/bin/fish", "/bin/zsh", "/bin/bash", "/bin/sh"],
    "login": false
  }
}

Browse 50 resource views across 9 sidebar groups — workloads, network, config, storage, Helm, Flux, RBAC, diagnostics, and custom resources. The sidebar auto-collapses to show only the active group.

Operate without leaving the terminal — exec into pods, launch ephemeral debug containers, open guarded node debug shells, port-forward, scale, restart, pause/resume/undo rollouts, rollback Helm releases, delete, edit YAML in $EDITOR, trigger CronJobs, cordon/drain nodes, reconcile Flux resources, and apply built-in resource templates.

Inspect with a bottom workbench that holds persistent tabs — YAML, drift diffs, rollout control, Helm history, decoded Secrets, timelines, pod/workload logs, exec sessions, extension output, AI analysis, port-forwards, relationship graphs, NetworkPolicy analysis, pod reachability, and traffic debugging.

Point and click where it helps — mouse capture is enabled in the TUI, with wheel scrolling for the pane under the pointer, sidebar row activation, table row selection/open, filter-bar focus, secondary-pane focus, workbench tab selection, and workbench border drag-resize.

Monitor via a dashboard with health gauges, CPU/memory utilization, namespace breakdown, top consumers, alerts, a Health Report, sanitizer findings, and Trivy-backed vulnerability summaries. Live watch streams keep 10 core resources updated in real-time.

• Rollout Control Center with revision-aware inspection, pause/resume, restart, and undo for Deployments, StatefulSets, and DaemonSets
• Helm release history, computed-values diff, and in-app rollback confirmation
• Advanced log investigation with presets, regex mode, time windows, jump-to-time, structured JSON summaries, and request correlation
• Workspaces, banks, and configurable hotkeys for repeatable operator layouts
• Health Report, sanitizer findings, vulnerability center, NetworkPolicy analysis, traffic debug, and AI-assisted analysis workflows
• Resource drift view, ephemeral pod debugging, node debug shell, and built-in apply templates
• Mouse support for everyday navigation: scroll the pane under the pointer, focus panes, choose rows, open selected rows, edit the filter bar, switch workbench tabs, drag-copy Pods names/rows via OSC 52 clipboard, and drag-resize the workbench
• Stability sweep for safer YAML handling, cleaner terminal recovery on startup failures, and safer $EDITOR handoff when editing resources

Press ? in-app for the full list. Summary:

Press Enter on a resource, then:

Exec tabs reserve text input for the remote shell. Press Esc once to enter controls mode, then use z, ,/., Ctrl+W, r, t, y, or S; press i or Enter to return to shell input. To launch exec in a native terminal, configure exec.external_terminal_template, for example:

{
  "exec": {
    "external_terminal_template": "kitty kubectl {context_arg} exec -it -n {namespace} {pod} -c {container} -- {shell}"
  }
}

kubectui [OPTIONS]

  --theme <name>          dark | nord | dracula | catppuccin | light
  --profile-render        Enable render profiling
  --profile-output <dir>  Profile output directory
  -V, --version           Version
  -h, --help              Help

• : discovers all available actions for the current resource — no need to memorize keys
• / search is live — filters as you type, Ctrl+U clears
• Sidebar groups auto-collapse — only the active group is expanded. Enter on a header temporarily expands it
• l on a Deployment/StatefulSet aggregates logs from all its pods
• P in log viewer shows previous logs from crashed containers
• W, {, and } let you save and cycle named workspaces for common operator flows
• AI workflows are opt-in and palette-driven; configure them in kubectui-config.json
• Vulnerability data appears when Trivy Operator CRDs are available in the cluster
• CPU/memory metrics require metrics-server in the cluster
• Data auto-refreshes every 30s (configurable via refresh_interval_secs in config)

MIT