Licensing
Licensing Noetis engines
Which arrangement fits you
Internal company team
Use an engine in your own systems through your account and key. Usage is metered per run, with per key monthly and per minute limits and an account wide credit hold.
SaaS company
Your customers can see engine outputs inside your application under a written agreement covering presentation, support and data handling. Discuss before launch.
Agency
Building for separate clients needs an agreed account structure so usage, data and billing stay separated per client. There is no standard agency licence yet.
White label or OEM
Outputs carry no Encevra branding today. Contractual white label, redistribution, support levels and private deployment are by agreement only.
Rights at a glance
| Arrangement | State | Notes |
|---|---|---|
| Call the API from your own servers with your own key | Available today | Keys are issued per account. Live use requires an approved account. |
| Use engine outputs inside your internal tools | Available today | Within your account's usage and the Terms. |
| Show engine outputs to your own customers in your product | Discuss requirements | Depends on a written agreement covering presentation and support. |
| Build for separate client companies | Discuss requirements | Account structure and billing per client are agreed case by case. |
| Remove Encevra branding (white label) | Discuss requirements | Outputs carry no Encevra branding today; contractual white label terms are agreed separately. |
| Resell or redistribute API access | Not offered | Keys must stay on your servers and are not transferable. |
| Private hosting or in region hosting | Discuss requirements | Not technically available today. Requirements are discussed, not promised. |
| Bring your own model provider key | Not offered | Not supported by the current runtime. |
| Ownership of model weights | Not offered | Noetis runs on third party language models. No weights are licensed or transferred. |
Nothing on this page grants rights by itself. The Terms and any signed agreement govern.
Practical questions
Usage measurement. Every run reports tokens and model cost. Self serve accounts are charged only for released runs: twice the measured model cost of the run, minimum $0.01, never more than the hold.
Failure billing. Incomplete, failed and cancelled runs are not charged under self serve billing.
Models. What you license is the engine workflow. It runs on models from OpenAI, Anthropic and Google DeepMind, and which ones depends on the engine. Per request provider choice is not available; a specific arrangement can be written into an agreement.
Model providers we build with
Data handling. Material is processed by third party model providers to produce the run. See Security and Trust for current facts.
Evaluation. Scoped evaluations on your own representative tasks are arranged on request. See the evaluation checklist.
Prices. See pricing.