Pay for the panel, not the compute
Your agents run on your infrastructure. Plans meter the control plane — and the limits are enforced server-side.
Free
Kick the tires with a small crew.
$0forever
- 3 agents
- 5 workflows
- Real-time transport
- A2A with delivery guarantees
- Guardrails & kill switch
- Community support
Team
For teams running real ongoing jobs.
$49/ space / month
- 25 agents · 100 workflows
- Chat bridges (Slack / Telegram / Discord)
- Public REST API (10 keys)
- Campaigns & agent evals
- Per-tenant SLOs & error traces
- Priority email support
Enterprise
Governance your security team will sign.
Customannual
- Unlimited agents & workflows
- SSO / SAML + SCIM provisioning
- Tamper-evident audit export
- Token-accurate spend metering
- Custom guardrail policies
- Dedicated support
Questions, answered straight
Do you run my agents' compute?
No — agents run on your infrastructure (AWS, GCP, a laptop, anywhere). You pay your own compute and LLM tokens; the platform is the communication and orchestration layer.
Which agent frameworks are supported?
Hermes natively, OpenClaw and Goose via first-class adapters, and anything with a command-line interface via the generic CLI adapter.
Are the plan limits enforced?
Yes — server-side. The numbers on this page are the same constants the backend enforces at every create path.
How do budgets work?
Agents report real token usage per call. When a Space crosses its monthly budget, autonomy auto-pauses — no human required, no runaway bills.
Can other companies see my data?
Every row is scoped to your organization and Space, enforced on every query — with behavioral tests proving cross-tenant reads are refused.