Galileo Open-Sources Agent Control to Enforce AI Policy at Scale
As enterprises deploy growing fleets of AI agents, keeping them consistent, compliant, and safe has become a serious engineering challenge. Galileo, the AI observability company, tackled this head-on today by releasing Agent Control — an open source control plane for governing AI agents at scale.
What It Does
Agent Control works as a centralized policy layer sitting above individual agents. Instead of writing safety rules, behavior guardrails, and compliance checks separately for each agent or framework, developers define policies once and they're enforced everywhere. The platform addresses a common enterprise headache: agents built on different frameworks behaving inconsistently, especially in sensitive business processes.
Apache 2.0, No Vendor Lock-In
Galileo released Agent Control under the Apache 2.0 license. Vikram Chatterji, Galileo's co-founder and CEO, explained: "Agent Control lets developers define guardrails once and apply them everywhere. By open-sourcing this under Apache 2.0, we're ensuring every enterprise and developer community can use it without vendor lock-in."
Day-One Integrations
Four platforms are integrating with Agent Control from launch: Strands Agents, CrewAI, Glean, and Cisco AI Defense. The mix covers agent orchestration, enterprise knowledge management, and network security — signaling broad adoption targets across IT departments.
Why It Matters
Enterprise AI agent adoption is accelerating faster than governance tooling. Agent Control — with cross-framework support and permissive open licensing — gives organizations an auditable, consistent way to enforce policies without rebuilding guardrails for every new agent deployment. Galileo, already known for AI observability, extends its platform into runtime policy enforcement with this release.