Pricing

One enterprise agreement, priced to the processes you run

No seat licences and no per-message fees. You pay for processes that are live and the volume they execute, agreed annually before anything ships.

Enterprise · the only plan · deployed in your own cloud

Annual

Custom — scoped per process

We scope the first process with you before quoting, so the number reflects the work rather than a tier you have to grow into.

Book a scoping call Talk to sales

What pricing is based on

Processes in production TIERED

Counted, not seats.

Monthly execution volume METERED

Runs, not messages.

Deployment model FLAT

Your cloud or ours.

Onboarding & support FLAT

Named engineer.

Deployment readiness

You see the same readiness board we do

Before a process goes live, the admin console shows exactly what is configured and what is missing. Nothing ships on our word alone.

Model providers and keys, per environment
Users, roles and access
Per-agent model assignment

What every agreement includes

No feature held back for a higher tier

The variables are scale and deployment, not capability.

Capability	Scope	Incl.
Conversational agents	Unlimited agents	✓
Deterministic workflows	Phases, steps, gates	✓
Execution analytics	Full run history	✓
Memory management	Per-process policies	✓
SSO, audit log, role controls	Org-wide	✓
Named onboarding engineer	Through first go-live	✓

Common questions

Is there a self-serve or starter plan?

Not yet. Every deployment currently goes through a scoping call because the first process is built with you, and that work does not fit a checkout page.

Can we run it in our own cloud?

Yes — that's the default. The platform is deployed in your own cloud, against your keys, and your data never leaves your account.

Which model providers are supported?

Providers and keys are configured per environment, so you choose which models each agent runs on. New OpenAI-compatible providers can be added without a platform change.

What does onboarding a process involve?

We scope one process with you, build and deploy it together in your environment, and stay close through first go-live — with a named engineer, not a support queue.