Bring human review into multi-agent teams

Human-in-the-loop is now available for Teams, with full support in the AgentOS chat interface and a dedicated API layer. Operators can review, intervene in, and steer team-level decisions the same way they already can with single agents.

Multi-agent teams often produce more consequential output than individual agents, since they coordinate across roles to complete higher-stakes tasks. Adding HITL at the team level closes a governance gap for organizations deploying teams in customer-facing or regulated workflows.

This gives platform owners a consistent oversight model across single agents and teams, so review processes, escalation paths, and compliance controls work the same way regardless of how an agent system is structured.

Learn more about HITL for Teams in our docs.