Pricing
Sized by the team. Never by the number of hosts.
Every plan covers an unlimited number of hosts and the whole product: dry runs, approval gates, roles, the audit log. Plans differ by how many people sign in and by what security and procurement need around them. Pricing is quoted per organisation.
Team
Up to 10 users
One team that owns its hosts and wants every change dry-run and on the record.
- Unlimited hosts
- Plans drafted from a request, with your own model key or ours
- Dry run, and the diff between two runs
- Approval gates on protected environments
- Roles, and access grants that expire on their own
- Scheduler with overlap protection
- Append-only audit log
- Email support
- Billed annually
- Priced per organisation
- Unlimited hosts
Business
10 to 50 users
Several teams sharing one inventory, with security and audit looking in.
- Everything in Team
- SAML and OIDC single sign-on
- Vault and Infisical as secret providers
- Plans synced from git, and dynamic inventory
- Audit log export
- Priority support
- Billed annually
- Priced per organisation
- Unlimited hosts
Enterprise
Unlimited users
The whole organisation, with the terms procurement needs.
- Everything in Business
- Customer-hosted execution: credentials never leave your network
- EU or US data residency, contracted
- Security review, DPA and subprocessor list
- Dedicated support with a named contact
- Annual invoicing and custom terms
- Billed annually or on invoice
- Priced per organisation
- Unlimited hosts
The AI composer works with your own model key on every plan.
Questions
What people ask before the quote.
How is Opafra priced?
Per organisation, by plan, billed annually. Each plan covers a band of users and an unlimited number of hosts. Tell us the size of the team and we quote against it.
Why is there no starting price on this page?
Because a starting price would be a number for the smallest possible team, and almost nobody is that team. What changes a quote is the band of users, whether execution runs in your network, and where the data sits. Those are three questions, and we would rather answer them against your setup than publish a figure you would have to correct on the first call. Tell us the size of the team and what it runs, and the quote comes back against that.
Do you charge per host?
No. Every plan covers an unlimited number of hosts. Every other option in this market bills per managed node, which is why fleets stop being registered in the tool that is supposed to govern them.
How does the AI composer work on the bill?
Two ways. Bring your own key from your model provider and the composer runs against your account, with nothing added to the bill. Or use ours, with a monthly drafting allowance sized to the plan. Plans you write by hand, sync from git or take from the module library never count either way.
What counts as a user?
A person who can sign in. Viewers count; hosts, environments and scheduled runs do not. Moving between bands is a conversation, not a lockout.
Where does the data live?
You choose EU or US at setup, and both cost the same. The security page describes the architecture and the subprocessors. Enterprise can run execution inside its own network, so credentials never leave it.
Can I cancel?
Yes. You keep access through the period you have paid for, and on the plans that include export your audit history stays exportable.
Get a quote against your team.
Tell us the size of the team and what it runs.