Zentinelle AI
Agent observability, policy evaluation and evidence for connected AI workloads.
Zentinelle AI records activity and evaluates policy for agents connected through its supported hooks, proxy, gateway or SDK integrations. Coverage depends on the integration and configured enforcement points.
Teams can manage model access, tool permissions, usage limits and retained evidence. Compliance mappings help organize evidence for review; they do not certify a deployment or replace the customer’s assessment.
The Zentinelle AI core is MIT-licensed. Support, policy integration and implementation are available separately. It can be adopted independently of Calliope AI Workbench and Astrolift AI.
Product site: zentinelle.ai. Current company compliance status: Trust Center.