Terminal¶

The Terminal is a real browser PTY into your cluster context — kubectl, k9s and helm pre-installed, no local tooling required.

What it does¶

Full xterm.js terminal with a pre-configured KUBECONFIG for the current cluster
Multiple concurrent session tabs, each an isolated bash subprocess
Claude Code is not pre-installed — it is proprietary and not ours to redistribute. Run claude-install once to install it into your own home directory (it persists there), then claude to sign in with your Anthropic account or an exported API key. Cluster credentials are never sent to Anthropic unless you explicitly ask Claude to read something
Short-drop reconnect: a refresh or network blip reconnects to the same session and replays the output buffer, so you keep shell state and scrollback — for about two minutes (TERMINAL_SESSION_TTL, default 120s). Past that the session is reaped and you start a new one. Opening the Terminal tab does not spawn a shell; starting a session is explicit, and a session belongs to one account and one browser tab
Per-session memory gauge. Memory is bounded by the terminal container's limit (terminal.resources.limits.memory, default 8Gi), not per session — the old per-session RLIMIT_AS cap was removed because it limited virtual address space and broke every Go binary in the image, kubectl, helm and k9s included

Click Terminal in the sidebar (AI group). Open additional tabs with the + button in the session bar.

Run kubectl, helm, k9s, or any standard Linux tool against your cluster
Launch claude for an interactive AI coding/ops session inside the pod
Copy-on-select — highlighting text copies it to your clipboard automatically
Expert Mode (expertMode.enabled=true) elevates the terminal's Kubernetes credentials to cluster-admin for write operations (kubectl delete, apply, exec) — the UI marks this clearly when active

A walkthrough of session tabs, the Claude Code workflow, and Expert Mode indicators will be added here in a follow-up pass.