Pricing

Pricing that scales with your clusters.

Start free on a single cluster. Turn on autonomous remediation and cost optimization when you're ready. Bring SSO, audit, and self-hosting for the whole fleet. You only pay for what Capnode actually watches — measured per vCPU under management, never per seat.

Deploy Agent Compare plans

Choose a plan

One agent, three ways to run it

Every plan installs the same RBAC-scoped, least-privilege Capnode agent into your cluster. The difference is how far Capnode is allowed to act — from read-only triage to fully autonomous, human-in-the-loop healing across your fleet.

Free

For a single cluster and engineers who want to see Capnode detect and explain — before it acts.

Free

No card required. Self-serve install.

One Kubernetes cluster
Core incident detection — 25+ failure modes, no Prometheus required
Ask Aria — read-only diagnosis and evidence-backed answers
Live topology and cluster state
Community support
Autonomous remediation
Cost optimization

Start free

Team

For teams running production across several clusters who want Capnode to heal and right-size automatically.

Let's talk

Usage-based, per vCPU under management.

Everything in Free, across multiple clusters
Autonomous remediation — safe actions auto-heal, risky actions need a click
AI workloads management — continuous right-sizing of requests, limits, replicas
Cost optimization — dissolve idle dev resources at rest, restore on first push
Integrations — Slack, Jira, and webhook alerting
Standard support, business hours

Contact sales

Enterprise

For organizations that need fleet-wide control, governance, and the option to keep everything in their own environment.

Custom

Volume pricing and committed-use terms.

Everything in Team, unlimited clusters
SSO and granular role-based access control
Full audit log of every detection, decision, and action
AI security posture scanning across the fleet
Self-host option — run the Capnode server in your own environment
Dedicated support with an SLA

Contact sales

Compare capabilities

What's included in each plan

From read-only triage to autonomous, fleet-wide healing. Every tier runs the same least-privilege agent — the table shows how far Capnode is allowed to act.

Capability	Free	Team	Enterprise
Clusters	1	Multiple	Unlimited
Incident detection (25+ failure modes)	Included	Included	Included
Ask Aria — AI diagnosis	Read-only	Full	Full
Autonomous remediation (HITL-gated)	—	Included	Included
AI workloads management	—	Included	Included
Cost optimization	—	Included	Included
Integrations (Slack, Jira, webhooks)	—	Included	Included
AI security posture scanning	—	Basic	Fleet-wide
SSO & role-based access control	—	—	Included
Audit log	—	Action history	Full audit
Self-host option	—	—	Available
Support	Community	Standard	Dedicated + SLA

Pricing FAQ

Questions about plans & billing

How Capnode is metered, where it runs, and what's protected by design. Still unsure? Talk to us.

Capnode is metered per vCPU under management — the compute your agent is actually watching across your clusters — not per user or per node. There are no seat licenses, so adding engineers never costs more. The Free plan is genuinely free for a single cluster. Paid usage scales smoothly with your fleet, and we'll walk you through the exact rate for your footprint when you reach out.

In both cases the Capnode agent runs inside your cluster — it streams state to a Capnode server over a long-lived, outbound WebSocket. On the hosted plans we operate that server for you. With the Enterprise self-host option you run the server in your own environment, so cluster data never leaves your network. The agent and the product behave identically either way; self-host is about where the control plane lives, not what Capnode can do.

A cluster is one Kubernetes cluster with a single Capnode agent installed, regardless of how many namespaces, nodes, or workloads it contains. The Free plan covers one cluster; Team covers several; Enterprise is unlimited. Because paid pricing is metered on vCPU under management rather than cluster count, a few large clusters and many small ones are billed on the same fair, consumption-based basis.

Free includes community support and full product documentation. Team adds standard support during business hours for installation, configuration, and remediation tuning. Enterprise includes dedicated support with a defined SLA and named contacts. Every plan ships with the same self-healing engine — higher tiers add people and response guarantees, not core capability.

The agent is RBAC-scoped and least-privilege, with a contained blast radius by design — it never mutates its own namespace. Remediation is tiered: safe actions run automatically, while risky actions always require a human click, so true human-in-the-loop control is built into every paid plan. Enterprise adds SSO, granular RBAC, and a full audit log of every detection, decision, and action. See the security architecture for the full model.

Yes. Start on Free, upgrade to Team when you want autonomous remediation and cost optimization, and move to Enterprise when you need SSO, audit, and self-hosting. Because billing is usage-based, your cost follows the vCPU you actually have under management — scale down and it scales down with you. There are no long-term lock-ins on the standard plans; Enterprise committed-use terms are agreed up front.

Heal your first cluster this afternoon

Install the Capnode agent, watch it detect and explain in minutes, then turn on autonomous remediation when you're ready. We'll size the right plan for your fleet.

Deploy Agent Explore features