Operon prices by agents in production, not per run. Hosted AI comes as an included usage allowance sized to your estate, with no per-token billing inside it. Build, test, and stage as many agents as you like at no charge. Professional and Enterprise Hosted run on Operon Cloud; Enterprise Private Deployment runs inside your own boundary.
For teams shipping governed AI to production. Ten agents live in production. Build, test, and stage as many as you like for free. Hosted AI is included as a defined usage allowance — flat and predictable, with no per-token billing inside it. Add capacity with Agent Packs.
For organizations scaling governed AI across many teams. Seventy-five agents in production on Operon's managed cloud, all seven routing tiers including human-in-the-loop review, federation, and the compliance pack, with a larger included usage allowance and no per-token billing inside it.
For regulated estates and sovereign deployments. Operon runs inside your own boundary: your VPC or private region, embedded as a Snowflake or Databricks Native App, or fully air-gapped — with your own models. Unlimited production agents, priced by arrangement, your compute, your boundary.
You pay for agents in production, not for what you build and not per run, with no per-token billing inside your usage allowance. Here's exactly how we count.
One autonomous agent, a single identity with its own role, policies, memory, and audit trail. This is the unit we count.
A deployed automation that combines agents with deterministic steps, forms, processes, reports. A flow can run one agent or coordinate many.
A governed team: a Lead orchestrating Workers, with a Guard that can intervene and a Scout that observes. The production unit.
The deployment a Fleet runs inside: one dispatch bus and one knowledge graph. Part of every Fleet, not a separate thing you buy.
Two or more Fleets combining as peers into one composite fleet, under a union agreement, sharing knowledge without copying it.
So how does a Fleet fit inside 10 agents? Professional's ten production agents cover two or three Fleets, ten standalone agents, or any mix in between. Need more without a sales call? Agent Packs add production agents to Professional in increments. Scaling across many teams steps up to Enterprise Hosted (75 agents in production); federated estates across regions or legal entities, air-gapped deployments, and unlimited agent counts are a custom Enterprise Private Deployment arrangement.
| Feature | Professional | Enterprise Hosted |
Enterprise Private Deployment |
|---|---|---|---|
| Agents & Scale | |||
| Agents | 10 in production | 75 in production | Custom |
| Build / test / staging agents | Unlimited | Unlimited | Unlimited |
| Fleets & Agent Mesh (multi-agent) | ● | ● | ● |
| Federation (composite fleets) | — | ● | ● |
| Bring-your-own-keys / model | Optional | Optional | ● |
| Hosted AI (included allowance) | ● | ● (larger) | Your model |
| Human-in-the-loop (Tiers 6 & 7) | — | Add-on | Add-on |
| Execution & Scale Limits | |||
| Concurrent executions | 10 | 75 | Your infra |
| Data transfer (hosted) | 100 GB/mo | 1 TB/mo | Your infra |
| Workflow execution timeout | 4 hours | 12 hours | Unlimited |
| Execution history | 90 days | 1 year | Custom |
| Log retention | 90 days | Custom | Custom |
| At-limit behavior | Throttle + alert | Throttle + alert | — |
| Builder Plane | |||
| NL2Operon | ● | ● | ● |
| Visual canvas | ● | ● | ● |
| SKILL.md library | ● | ● | ● |
| Import pipeline | ● | ● | ● |
| Runtime Plane | |||
| Multi-tenant runtime | ● | ● | ● |
| Smart routing & fallback | ● | ● | ● |
| Federation across boundaries | — | ● | ● |
| Control Plane | |||
| OPA/Rego policy enforcement | ● | ● | ● |
| EDT to warehouse | ● | ● | ● |
| Compliance pack (SOX/HIPAA/GDPR/EU AI Act) | — | ● | ● |
| Scoped emergency pause | ● | ● | ● |
| Deployment | |||
| Operon Cloud (managed) | ● | ● | — |
| Snowflake / Databricks Native App | — | — | ● |
| Private deployment (VPC, Native App, air-gapped) | — | — | ● |
| Identity & Support | |||
| SSO (SAML / OIDC) | ● | ● | ● |
| ABAC / RBAC | ● | ● | ● |
| Standard SLA | ● | — | — |
| Premium SLA + named support | — | ● | ● |
| Architecture review session | — | ● | ● |
A clear, editable model of what Operon is worth in your environment. Seven value categories. Conservative defaults. No login required.
Run the ROI numbers →Ten agents live in production. You can draft, test, simulate, and stage unlimited agents for free, an agent only counts once you promote it to production. Pause a production agent and it releases its slot. So a team might build fifty agents over a quarter and run ten; you pay for the ten.
No. An agentic flow is the deployed automation. It can run a single agent or coordinate many, alongside deterministic steps like forms and processes. We count agents (identities), not flows. A flow that runs one agent counts as one.
By member agent. A Fleet with a Lead and three Workers uses four of your allowance. The Hive the Fleet runs in is part of the Fleet, not a separate line item, and Federation joins existing Fleets rather than adding agents. Professional's ten production agents cover two or three Fleets or ten standalone agents; more than that is an Enterprise conversation, and Agent Packs add capacity in increments in between.
Hosted editions ship defined envelopes: Professional includes 100 GB/month of data transfer, 10 concurrent executions, 4-hour run timeouts, and 90-day history; Enterprise Hosted raises these to 1 TB/month, 75 concurrent, 12-hour timeouts, and a year of history. Hit a ceiling and we throttle and alert rather than hard-stop or surprise-bill. Overage is opt-in. Private deployments run on your infrastructure, so you own the ceilings. Heavy analytical work is pushed to run in your connected warehouse, so scanning large datasets doesn't burn your transfer allowance.
Enterprise Hosted runs on Operon's managed cloud with a flat $2,500/month for 75 production agents and a defined scale envelope. Enterprise Private Deployment is for regulated or sovereign estates that run Operon on their own infrastructure — in your VPC or private region, embedded as a Snowflake or Databricks Native App, or fully air-gapped — with your own models. Air gap is one option within it, not the whole tier: a Native App runs inside your account but is still networked, while air-gapped means nothing leaves at all. Because you own the compute, it's priced by arrangement rather than metered, and production agents are unlimited.
Yes. Agent definitions, policies, and EDT history all carry forward. Moving from Professional to Enterprise, or from hosted to your own private cloud, is a configuration change, not a migration.
A 45-minute session with our principal architects for Enterprise prospects. We'll bring a tailored architectural sketch of how Operon would compose with your platform mix, and a buy/build evaluation rubric. Run by architects, not sales engineers.
With a 45-minute architecture review run by our principal architects. The deliverable is a tailored sketch of how Operon composes with your current platform mix, plus a buy/build rubric you can take to your own committee. Nothing about it is a sales demo, and you keep the output whether or not you proceed.
Tell us which edition fits your estate and we'll take it from there.
Tell us your organization, roughly how many agents you expect to run in production, and whether you need hosted or your own cloud.