Decisions that fit your work.
Build a narrow decision into an existing workflow. Keep your application in control of what happens next.
A choice, a score, or a check
Route a support case, score an application against an explicit rubric, or check whether a condition holds. Give Ren 1.1 the situation and the question; get structured probabilities your code can use.
The server returns a gate action for the threshold you choose. Send uncertain cases to someone who can decide. The gate does not turn model confidence into a guarantee.
Your data stays yours
We do not train on customer inputs, outputs or contact messages. API request content is used to answer your request, not to build a training set. Usage records are kept separately from request content.
Discuss a zero-retention arrangement before sending sensitive data. It must cover the gateway, inference service and operational logs. It is an option to agree in writing, not a setting enabled by this page.
Agree on handling before you integrate
We can work through a data processing agreement, access controls, retention and deletion requirements before access is enabled. Ask us which service providers and processing locations apply to your account.
We do not claim a security certification, a dedicated region, an on-premises deployment or a regulated-data approval that we have not supplied in writing.
Start with your own decisions
Bring a representative set of situations, clear options and examples of mistakes that matter. Evaluate outside production, set a review threshold and monitor disagreements. Keep consequential decisions under human review.
Contact us about private beta access, account limits or an enterprise data handling review.