Why We Built a CLI for FinOps (And Why GUIs Fail Engineers)

FinOps tools have a usability problem.

They are almost always massive, complex web dashboards. They have beautiful pie charts, executive summaries, and PDF export buttons. They are built for CFOs.

But CFOs don’t fix infrastructure. Engineers do.

The Context Switch Tax

Imagine this workflow:

1. You’re debugging a deployment in your terminal (kubectl get pods).
2. You wonder, “Is this over-provisioned?”
3. You open a browser tab.
4. Log in to CloudHealth / Kubecost / AWS Cost Explorer.
5. Navigate 4 menus deep to find the right cluster and namespace.
6. Filter by the last hour.
7. …By now, you’ve forgotten what you were debugging.

Cost data is context. It shouldn’t be a destination; it should be an overlay on your existing work.

The ClusterCost Philosophy: kubectl cost

We built ClusterCost to be CLI-first because that’s where the work happens.

We believe:

1. Speed is a Feature: You should get a cost estimate in <500ms, not 5 seconds of loading spinners.
2. Text > Charts: For an engineer, a precise table of $/hour is more actionable than a colorful donut chart.
3. Integration is Key: A CLI tool can be piped into scripts, CI/CD pipelines, and Slack bots. A web dashboard is an island.

Empowering the Edge

The goal isn’t to make engineers “think about money” every second. It’s to give them feedback loops.

If running kubectl apply showed you the projected cost increase immediately, you’d make better decisions naturally. No meetings required.

That’s why we built the Local Dashboard and the CLI Agent. To bring the data down from the ivory tower and put it in the hands of the people who actually build the cloud.

Try it yourself

No signup required. Just run:

npx clustercost@cli

See your cluster’s cost in seconds. Then get back to work.