Overview

RaiCentral (RAI Control Plane) is a runtime governance layer for AI agents. It sits between your agents and the business systems they act on, and it enforces policy before a consequential action proceeds.

Alpha status: RaiCentral is in alpha. Pilots run under agreed terms. There is no hosted multi-tenant signup for Community Edition — you run the instance you operate. We do not yet hold third-party certifications.

What RaiCentral does: registers AI systems, agents, versions, and owners; authorizes tool calls and consequential actions against policy; enforces allow, deny, and human-approval decisions with fail-closed defaults; and keeps decision receipts linked to reported execution results and spend where you integrate metering.

Who it is for: platform engineers integrating governance into agent applications; governance and security teams that need runtime enforcement and reviewable evidence; teams running scoped pilots on one consequential workflow.

Core principle: governance must happen at runtime, not after the fact. Your application asks for a decision before a side effect, executes only on ALLOW, then reports what happened. RaiCentral governs actions your application routes through it — it does not intercept traffic automatically.

Next steps: follow the Python or TypeScript quick start, then use Pilot onboarding to stand up a Community Edition instance and issue a runtime agent key.