Priced on what runs, not on who logs in.
Every deployment is quoted. Unlimited users and unlimited workflows wherever it runs; usage is counted in full workflow executions plus AI spend, never per step.
OCNORA Cloud
Priced per deployment in one call.
- Hosted and operated by us
- Executions and AI budget set in your quote
- Templates
- Email support
Your cloud
Priced per deployment in one call.
- Offline signed licence, no phone-home
- All governance features
- Helm for AWS, Azure and GCP
- Your own model provider keys
- Email support
Embedded
Priced per deployment in one call.
- Per-tenant licences
- JIT SSO for your users
- Per-tenant plugins and images
- Partner support
Licence model: a subscription for OCNORA Cloud, a per-deployment signed licence for your cloud, and a per-tenant licence for embedded partners. Enterprise options on any deployment: dedicated support and an SLA, a security review under NDA, design-partner roadmap access and custom limits.
Four things set the number.
Executions per month
Full workflow runs, from trigger to finish. Never counted per step, per user or per workflow.
AI spend
Whether AI is involved and how much. On OCNORA Cloud an AI budget is part of the quote; on your cloud you bring your own provider keys.
Where it runs
OCNORA Cloud, your own cloud, or embedded in your product. For embedded, how many tenants you serve.
Support
Email, a partner channel for embedded, or dedicated support with an SLA and a security review under NDA.
Running an enterprise evaluation? Plan it with the evaluation guide.
The same product everywhere. The difference is where it runs and who supports it.
| Feature | OCNORA Cloud | Your cloud | Embedded |
|---|---|---|---|
| Build | |||
Unlimited users and workflows | |||
Visual builder, expressions, test step | |||
Build with AI | |||
n8n JSON import | |||
Templates | |||
| Agents | |||
Agent node, tools, memory | |||
Hosted models: Anthropic, OpenAI, Google and DeepSeek | |||
Local models through any OpenAI-compatible inference server you host (vLLM, Ollama, LM Studio, llama.cpp) Needs your own inference host | |||
MCP server and client | |||
| Govern | |||
AI budgets and circuit breakers | |||
Human-in-the-loop approvals | |||
Audit log on every run | |||
Per-tenant budgets | |||
PII crypto-shred | |||
Dashboard sign-in: email and password, MFA, OIDC or SAML 2.0 single sign-on, SCIM 2.0 provisioning OIDC and SP-initiated SAML 2.0 where your installation has an identity provider configured; SCIM 2.0 provisioning from it. Tenant-key SSO for embed hosts. | |||
| Run | |||
Workflow Health and Auto Fix | |||
Ops center and business reports | |||
Error workflows, retries, timeouts | |||
| Deploy | |||
Hosted by OCNORA | |||
Docker Compose and Helm | |||
Offline licence | |||
Embed by iframe with JIT SSO | |||
Per-tenant plugins and images | |||
White-label | |||
| Support | |||
Email support | |||
Partner support channel | |||
Dedicated support and SLA Offered on any deployment as part of an Enterprise quote. | partial | partial | partial |
Security review under NDA Offered on any deployment as part of an Enterprise quote. | partial | partial | partial |
Design-partner roadmap access Offered on any deployment as part of an Enterprise quote. | partial | partial | partial |
partial marks a capability with a caveat or offered by quote; the note in the row says which.
Questions we get on the first call.
Why is there no price list?
We do not publish list prices yet. Every deployment is quoted from four things: executions per month, AI spend, where it runs and the support you need. One call is usually enough to give you a number.
What counts as an execution?
One full run of one workflow, from trigger to finish, whatever happens inside it. A workflow with forty steps and a loop is one execution. Sub-workflows called from a parent count as part of the parent’s run. Test runs from the builder do not count.
How is AI spend budgeted and capped?
Every model call is written to a cost ledger with its tokens and dollar cost. You set a monthly budget per workspace and, on Embedded, per tenant and per provider. When a budget is reached, a circuit breaker stops further AI calls and the run takes its fallback path; nothing keeps spending in the background. On OCNORA Cloud the AI budget is part of your quote; on your own cloud you bring your own provider keys and the same budgets apply.
Can I self-host?
Yes. OCNORA ships as Docker Compose for a single host and as a Helm chart with production values for AWS, Azure and Google Cloud. The stack runs as separate API, router, scheduler and worker roles on Redis queues, so it scales the way your cluster does.
Does the licence phone home?
No. A self-hosted or embedded licence is a file signed with Ed25519. Your installation verifies the signature with a public key and reads the limits from the file; it never calls back to us. When a licence expires there is a grace window before features are restricted, so a missed renewal email does not stop production.
Can I import my n8n workflows?
Yes. Export a workflow from n8n as JSON and drop it into the importer. About 45 n8n node types are mapped to OCNORA nodes; anything unmapped is imported as a placeholder step and flagged by Workflow Health so you can see exactly what needs attention before the first run.
What happens when I exceed my limits?
We talk. Limits are set with you in the quote, and going past them starts with a message from us, not a hard stop. On OCNORA Cloud, runs continue and we agree a new quote; on your cloud and embedded, the licence file sets the ceiling and we issue a new one.
How do I get a quote?
Tell us what you run: rough executions per month, whether AI is involved, and whether you need OCNORA Cloud, your own cloud or embedded. Use the contact form or write to sales@kxntech.com.
Tell us what you run. We'll price it in one call.
Executions per month, whether AI is involved, and where it should run. That is all we need to give you a number.