Pricing

Free where it counts

The kprompt CLI is open source and free forever. No seat limits, no usage meters, no credit card. Bring your own LLM key or run locally — you own the runtime and your data.

Free forever

The CLI stays free. Forever. No asterisk.

We build kprompt for operators, in the open. The core CLI — every command, every cluster, every run — is Apache-2.0 and will never sit behind a paywall. Install it, self-host it, fork it. It is yours.

Install in one line

No API key · kind cluster · ~60 seconds

Plans

CLI

Free forever
$0free forever

The open-source runtime. Yours to keep, always.

  • Full CLI — every command, no feature gates
  • Unlimited clusters, contexts, and runs
  • Bring your own LLM key (BYOK) or run local Ollama
  • Plan → approve → apply safety on every action
  • AI SRE: investigate, why, timeline, impact
  • Observe agent for continuous correlation
  • Works offline for the zero-LLM walkthrough
  • Apache-2.0 licensed — audit, fork, self-host

Team

Custom

Shared clusters, shared history, shared guardrails.

  • Everything in the free CLI
  • Shared team console at app.kprompt.ai
  • Run history and audit trail across the team
  • Role-based approvals for apply gates
  • SSO and centralized policy (coming soon)

Pricing FAQ

Is the CLI really free forever?

Yes. The kprompt CLI is open source under Apache-2.0 and free forever — no seat limits, no usage meters, no credit card. We will never paywall the core CLI.

What does the LLM cost, then?

kprompt is bring-your-own-key. You pay your provider directly (or $0 with local Ollama and the zero-LLM walkthrough). We never mark up tokens or sit between you and your model.

Do you see my clusters or data?

No. The CLI runs on your machine against your kubeconfig. Your cluster data and prompts go to the LLM provider you choose — not to us.

What is the Team plan for?

The optional Team console adds shared run history, audit trails, and role-based approvals for organizations. The CLI stays free regardless.

Nothing to buy. Just install and run.

Start with the zero-LLM walkthrough on a kind cluster — $0, no key. Level up with your own model whenever you want.