Know what your Kubernetes actually costs
See spend by cluster, by team and by node - and exactly how much is sitting idle. No agents, no Prometheus, no FinOps platform to stand up. Just connect your kubeconfig and watch where the money goes. A fast, native Mac app. Free for 14 days.
Works with Google GKE, Amazon EKS, Azure AKS, DigitalOcean, self-managed and more · macOS Sonoma (14) or later
The problem
Cloud spend is up. Nobody can say why.
Kubernetes makes it easy to ship and easy to over-provision. The cost shows up on the invoice long after the decision that caused it - and by then it's buried.
The bill grows, the answer doesn't
Kubernetes is the biggest line on the cloud invoice and the hardest to explain. “Why did it go up?” has no quick answer when spend is buried across nodes, namespaces and teams.
You're paying for capacity nobody uses
Teams request far more CPU and memory than they run. That headroom is provisioned, scheduled and billed - idle capacity you pay for every hour of every day.
No cost per team or product
Without per-namespace attribution there's no showback, no accountability, and no way to tell which team or service is actually driving the spend.
FinOps tooling is a project of its own
Kubecost, Prometheus, agents on every cluster, dashboards to host, a per-cluster SaaS bill - standing up cost visibility shouldn't need its own platform team.
What Kubius shows you
The whole cost picture, in one window
Open the Cost & FinOps view and your spend is laid out the way you'd actually reason about it - totals, waste, and who's responsible.
Estimated monthly spend
Every node priced from its exact machine type and region, summed into a clear monthly figure - across one cluster or many, in any currency.
Idle & unallocated waste
The gap between capacity you pay for and capacity your pods actually request, surfaced as real dollars and a percentage - the number that turns “we might be wasting money” into “we're wasting this much.”
Cost by namespace - showback built in
Spend split across pods and rolled up per namespace, sorted by cost. Instant per-team, per-product, per-environment breakdown you can take to a planning meeting.
Cost by node type & region
See spend grouped by instance type and region - spot the expensive pools, the over-sized node groups, and the regions quietly costing more.
Right-sizing reclaim, quantified
Sample real CPU and memory, then see exactly what each workload would reclaim if right-sized - the concrete savings, per workload, ready to apply in place.
Your rates, not list prices
Bundled on-demand pricing per machine type and region out of the box, or drop in your own negotiated $/vCPU-hour and $/GiB-hour so the numbers match your real contract.
Nothing to deploy
FinOps visibility, without the FinOps project
Most Kubernetes cost tools are infrastructure in their own right - agents on every cluster, a Prometheus to feed them, dashboards to host, and a recurring bill. Kubius is a Mac app. You're looking at real numbers minutes after you open it.
- No agents, no Helm charts, no Prometheus to stand up
- No per-cluster SaaS subscription - a one-time licence
- Connects straight from your kubeconfig, over TLS
- Your cost and cluster data never leave your machine
- GKE, EKS, AKS, DigitalOcean, self-managed and more, side by side
- Lives in the client your engineers already use
In your first afternoon
From “we think we're overspending” to a plan
- 1
Connect and see the total
Pick a cluster from your kubeconfig. Kubius prices every node and shows the monthly spend - no setup, no waiting on a data pipeline.
- 2
Find the idle money
The idle / unallocated figure shows what you're paying for but not using - the single biggest, fastest lever on the bill.
- 3
Attribute it by team
The per-namespace breakdown turns one big number into a conversation each team can own - showback without a chargeback project.
- 4
Right-size and reclaim
Sample real usage, see what each workload would reclaim, and apply right-sized requests in place - the savings made concrete, not theoretical.
Simple, one-time pricing
Pay once. No per-cluster SaaS bill.
One-time licence from €49 - no subscription. 14-day free trial, no card required.
Free 14-day trial · macOS Sonoma (14) or later · No agents, nothing to install in-cluster
Questions leadership asks
Do we have to install anything in our clusters?
No. Kubius reads your existing kubeconfig and talks to the Kubernetes API directly - no agents, no Prometheus, no Helm charts, nothing to deploy or maintain. Right-sizing uses the standard metrics API (metrics-server) when it's present; cost estimates need nothing at all.
How accurate are the cost figures?
Spend is estimated from each workload's CPU and memory requests, priced per machine type and region from bundled on-demand rates - or your own $/vCPU-hour and $/GiB-hour. It doesn't include spot or committed-use discounts, so treat it as directional: ideal for spotting idle waste, comparing teams and sizing decisions - not a replacement for your billing invoice.
Does our cost or cluster data leave our machines?
No. Everything stays between your Mac and your clusters - resource data, costs and all. Nothing about your clusters is ever sent to Kubius. The only thing that reaches our servers is your account and licence (your email).
Can we see cost per team?
Yes. The per-namespace breakdown gives you showback by team, product or environment out of the box - sorted by spend, with idle capacity surfaced separately so you can see waste as well as usage.
Which clouds and clusters are supported?
Google GKE, Amazon EKS, Azure AKS, DigitalOcean, self-managed and more - any context in your kubeconfig, viewed side by side. The app runs on macOS Sonoma (14) or later, Apple Silicon or Intel.
Is it a subscription?
No. Kubius is a one-time licence with free updates within the major version, usable on up to 3 of your Macs - no per-cluster SaaS fees and no seat-based platform bill.