Pricing

Same platform. Pick your operator.

Every Jōkamachi cluster runs the same open-source software. What changes between plans is who watches the cluster — your operators, our operators, KAOS, or any combination of the three.

Free

€0self-hosted, forever

Open-source. Run by you.

• Full Jōkamachi platform under an OSI-approved licence
• Bootstrap, Helm charts, kubenix management, booklets
• Read access to community forums
• Public bug tracker — file issues without a support contract
• No telemetry, no licence server, no expiry

Get started

Managed

+10% running costs+ billable operator hours

Our operators run the cluster.

• Everything in Free
• Cloud-hosted dashboard: curated metrics and live cost overview
• Write access to community forums
• Automated provisioning, upgrades, backups — included
• Hourly rate set by your SLA tier (Basic → Platinum)
• No monthly retainer — pay only when humans intervene

Sign up

Agentic

+10% running costs+ billable operator hours+ KAOS tokens at pass-through

KAOS runs the cluster, humans stay in the loop.

• Everything in Managed
• KAOS attached to your cluster
• Guardrails you control, per incident class
• Tokens replace operator hours where the playbook is mature
• Decision log streamed to the dashboard
• Token spend visible live, capped on request

Pilot KAOS

Agentic is not a higher tier than Managed. It’s a different operator. Customers on Managed get human attention; customers on Agentic get agentic attention with humans on call. Pick whichever matches your risk appetite.

The rest of this page expands each plan and explains how billing works, so you can pick one without a sales call.

Free

Self-hosted. Open source. The same software we run.

• The full Jōkamachi platform under an OSI-approved licence.
• Bootstrap scripts, Helm charts, kubenix management, and the booklets that explain them.
• Read access to the community forums, where Jōkamachi operators and paid-plan customers solve problems in the open.
• Issues and merge requests welcome on git.jokamachi.systems.
• No telemetry phoning home, no license server, no “free for evaluation only” footnotes.

You bring the cloud account and the on-call rotation. We bring the code and the manual.

If you outgrow self-hosting, the Managed plan picks up your existing cluster — no re-platforming, no migration project.

Managed

Managed has two pricing components, deliberately separated so you only pay for the human time you actually use.

1. +10% of running costs — automated platform operations. Covers the work of keeping the managed services current, healthy, and observable: image rebuilds, version tracking, CVE patching, the dashboard, the audit log, the cost-overview pipeline. It’s flat, predictable, and unrelated to how often a human is on call.
2. Operator hours — billed at the hourly rate of your chosen SLA tier. No monthly retainer. You pay only when humans actually intervene: incident response, manual changes, migrations, custom engineering. The tier you pick (Basic, Bronze, Silver, Gold, or Platinum) sets both the response SLA and the hourly rate. Hours are tracked, invoiced transparently, and shown live in the dashboard so there are no surprise invoices. See Subscription plans.

Automated work is free across both components — provisioning, upgrades, monitoring, alerting, and automatic backups are part of the open-source platform and don’t consume operator hours.

You also get, on top of the Free plan:

• Cloud-hosted dashboard with curated cluster metrics, a live cost overview, audit log, and incident history. The Free platform ships the same raw metrics — Managed turns them into views your on-call can read under pressure, not a wall of Prometheus.
• Write access to the community forums, alongside Jōkamachi operators and other customers.

Tenancy: yours or ours

Managed Kubernetes runs on either your Hetzner account or ours. Pick whichever fits your billing and procurement reality.

Your tenancy

• You own the Hetzner account and pay Hetzner directly.
• +10% management fee attaches to your invoice.
• You provision additional Hetzner resources when the cluster needs more headroom — we tell you exactly what to add.

Jōkamachi tenancy

• We own the Hetzner account and run the cluster on it.
• The Hetzner invoice passes through to you with our +10% applied.
• We provision more resources when the cluster needs them, without waiting on you.

Either way, the +10% is the same and is calculated on the same running cost. The choice is about who pays the hosting bill and who lifts a finger when limits are reached — not about price.

Agentic is jōkamachi-tenancy only. KAOS is closed-source and lives on infrastructure we control.

Agentic

+10% of your running costs, plus token cost at pass-through. KAOS runs the cluster; humans stay in the loop.

• Everything in Managed, plus —
• The Kubernetes Agentic Operator Substrate (KAOS) attached to your cluster, observing metrics and events.
• Guardrails you control: KAOS can be advisory-only, scoped to specific namespaces, or granted apply rights with a human approval step. As the playbook for a class of incidents matures, you can promote KAOS from propose to approve to act on that class — keeping humans in the loop where the playbook is still settling, and out of the loop where the same incident has been resolved a hundred times the same way. The KAOS booklet covers the maturity tiers in detail.
• Tokens replace hours. Where Managed bills operator hours for manual interventions, Agentic resolves the same class of work in tokens — typically far cheaper, and visible live in the dashboard.
• Decision log streamed to the dashboard — every action KAOS proposes or takes, with the reasoning that led to it.

How tokens are billed

We bill the underlying model provider’s invoice through to you with no markup. A typical small cluster spends in the range of $X–$Y per month on KAOS tokens; the dashboard shows live spend and lets you cap it. Stub — replace $X–$Y once we have real numbers from production clusters.

Subscription plans

No monthly retainer. The tier you pick sets your response SLA and the hourly rate that applies when humans actually intervene. We operate clusters, not individual servers, so pricing is by the hour — not per server.

Every tier includes operation support, in-depth monitoring, automated alerts, live-chat and ticket support, and the +10% automated-platform operations baseline (dashboard, audit log, cost overview).

Bronze and above additionally include:

• Backup and recovery.
• A monthly bucket of free consultation hours.
• An explicit response-time SLA at the tier’s coverage window.

Basic is the no-commitment tier — discounted hourly rates, request-driven access, no on-call guarantee.

Consultation engagements

Outside of incident response, Jōkamachi operators are available for architecture review, capacity planning, migrations, and bespoke engineering work. Hours are billed at the same hourly rate as the SLA tier you’ve signed up for.

What “running costs” means

Your cloud provider’s invoice for the compute, storage, and network attached to the cluster — not your whole AWS/GCP/Azure account. We will either bill on top of an invoice you forward us, or run the cluster in an account we own and pass the invoice through. Either way the +10% is on cluster-attributable spend, not on your unrelated workloads.

Always included

• Automated provisioning, upgrades, monitoring, and backups (in the open-source platform — no operator hours required).
• A copy of every booklet, kept in sync with the platform you’re on.
• The right to take your cluster and run it yourself at any time — there is nothing in Managed or Agentic that isn’t also in Free.