● see it in action

Watch it work,
not just read about it.

All five recordings below are captured straight from kute --demo's built-in fake cluster — the exact same one-command experience you get, unedited.

incident walkthrough

From a clean namespace to root cause in one flow.

A clean production namespace, then all-namespaces reveals a CrashLoopBackOff pod. Pod detail and logs show the actual cause, delete asks for explicit confirmation, and the incident timeline correlates the crash to a rollout ten minutes earlier.

kute: an incident walkthrough — a clean production namespace, then all-namespaces reveals a CrashLoopBackOff pod, pod detail and logs show the actual cause, delete asks for explicit confirmation, and the timeline correlates the crash to a rollout 10 minutes earlier

Regenerated with scripts/record-demo.sh docs/assets/demo-all-namespaces.tape

goto palette

One palette jumps to anything.

Typing a kind's alias letter pins it to rank 1 — still requires ↵ to jump, never fires instantly. Bare g + ↵ alt-tabs between kinds the same way the namespace palette alt-tabs between namespaces. And the fuzzy corpus searches resource names too: typing a specific pod's name jumps straight to it, switching kind as a side effect.

kute: the goto palette pinning Deployments to rank 1 with the "d" alias, alt-tabbing back and forth between Pods and Deployments with no typing, then typing "cache" to jump straight to the cache-0 pod and open its detail

Regenerated with scripts/record-demo.sh docs/assets/demo-goto-palette.tape

namespace palette

Alt-tab between namespaces with no typing.

The same palette that jumps to any resource kind also toggles between your last two namespaces — bare n + ↵ switches back and forth with nothing typed — and recalls older ones from the RECENT row by number as the digit assignments shift with what you've visited most recently.

kute: the namespace palette alt-tabbing to ingress-nginx and back to default with no typing, then recalling production and argocd from the RECENT row by their assigned digit

Regenerated with scripts/record-demo.sh docs/assets/demo-namespace-palette.tape

everyday actions

Every action shows its command first.

One Deployment, one continuous flow: scale warns you it's HPA-managed before overriding a no-op, set image picks a tag straight from rollout history, resource limits nudge next to the live usage bars, and a label add/remove shows the exact kubectl command before it runs. Port-forward starts without suspending kute — the header's forward chip stays live while you keep browsing.

kute: scaling a HPA-managed Deployment shows a 'managed by hpa' warning instead of a plain command preview, set image picks a tag from rollout history, set resources nudges a cpu limit, adding then removing a label shows the exact kubectl label command before the removal runs, and starting a port-forward leaves the header's forward chip live while browsing continues

Regenerated with scripts/record-demo.sh docs/assets/demo-everyday-actions.tape

routing table

Ingress and Gateway API, resolved live.

An Ingress with a healthy route sits right next to one whose Service was never created — red inline, not a silent failure discovered later. An HTTPRoute's weighted canary split resolves both backends' live endpoint counts, and p jumps straight to the parent Gateway that accepted it — resolving Gateway API's split ownership both ways.

kute: an Ingress routing table showing a healthy / route next to a broken /admin route whose backend Service doesn't exist, an HTTPRoute's 90/10 weighted canary split with live endpoint counts, and jumping via p to the route's accepted parent Gateway

Regenerated with scripts/record-demo.sh docs/assets/demo-routing-table.tape

SINGLE BINARY

Install kute.

Explore the full Kute experience with kute --demo—a built-in fake cluster with no setup or configuration.

$curl -fsSL https://kute.dev/install.sh | sh

Or install with Homebrew

$brew install kute-dev/tap/kute
View on GitHub See what's inside