Your whole Kubernetes cluster.
One application.

Download for Mac See features

01 / Core

Everything you manage, locally

Live views of pods, jobs, and nodes, with one-click actions on every row.

Pods, Live

  • • Real-time health: ready count, status, restarts, and live CPU/memory from metrics-server.
  • • Find problems fast: filter by namespace, sort by CPU/memory/restarts, or flip on "issues only".
  • • Act in one click: logs, exec, port-forward, YAML, delete, and auto-diagnose.

Jobs

  • • Full picture: every job, including completed and lingering ones, with status and completion detail.
  • • Spot failures: see which jobs errored or never finished, right next to their pods.
  • • Drill down: jump from a job to its pod's logs, YAML, or diagnosis in one click.

Nodes

  • • Capacity vs. usage: allocatable CPU, memory, and ephemeral storage against live usage, per node.
  • • Health and role: Ready status and node role (control plane or worker).
  • • Packing view: pods-per-node against the pod limit, so you see crowding and headroom.

02 / FinOps

Cloud cost, where you manage it

See cloud spend next to the workloads driving it. No separate dashboard, no data leaving your machine.

Daily spend, in-app

Live cloud cost data in the same window as your workloads.

Cost by service

Break spend down by service and see exactly what's driving the bill.

Zero extra setup

Reuses the cloud credentials your kube context already uses. If the cluster loads, cost usually does too.

finops · cloud daily spend
Today $63.63
Compute$21.40
Database$12.50
Kubernetes$9.60
Storage$6.85
CDN$5.20
Data transfer$4.33
NoSQL$3.75

daily cloud spend, by service

03 / AI Copilot

Ask your cluster anything

  • Full cluster context

    Understands your workloads, metrics, events, and cost. No copy-pasting YAML into a chatbot.

  • Answers from live data

    Runs read-only kubectl (get, describe, logs, top) so answers come from your cluster, not guesses.

  • Read-only, bring your own key

    Every mutating command is refused. Plug in Gemini or OpenRouter. Free models work out of the box.

04 / Rightsizing

Recommend & apply

  • Based on real usage, not guesses

    Samples CPU/memory every 30s over a 6h window and sizes to p95 + peak, so spiky apps never get starved.

  • Clear verdicts

    OK / UPSIZE CPU / UPSIZE MEM / DOWNSIZE, each with a concrete suggested request.

  • One-click apply

    Patch the new request straight to the owning workload. No YAML editing.

05 / Why us

Why Clusterheart

One tool, nothing to deploy, and your credentials stay put.

All-in-one

  • • One application for everything: inspect, diagnose, track cost, chat with AI, and right-size, all in one place.
  • • No tool sprawl: stop stitching five tools together to run one cluster.

Nothing to deploy

  • • Just a desktop install: point it at your kubeconfig and you're running.
  • • Zero cluster footprint: no agents, operators, or CRDs inside your cluster.

Private by design

  • • Everything stays local: cluster and cloud calls run on your machine, never through us.
  • • We only gate access: our server decides who's allowed in. It never sees your cluster.

06 / FAQ

Straight answers,
no fluff.

Get Clusterheart

Install the desktop app and point it at your kubeconfig. That's the whole setup.

macOS