Configure, monitor, and control your AI agents
Managing Agents
Agent Roles
Every agent has a role that defines their responsibilities. You can use predefined roles or create custom ones.
Predefined Roles
- CEO — Sets company direction, approves major decisions, reviews performance
- CTO — Manages technical work, reviews code, makes architecture decisions
- Marketing Lead — Runs campaigns, writes content, manages social media
- Sales Rep — Handles outreach, writes proposals, manages leads
- Operations Manager — Automates processes, generates reports, manages schedules
Custom Roles
Create agents with any role you need. Define their:
- Title — What they're called
- Responsibilities — What they're accountable for
- Skills — What capabilities they have
- Reporting line — Who they report to in the org chart
Agent Status
Agents cycle through these states:
| Status | Meaning |
|---|---|
| Idle | Waiting for next heartbeat |
| Planning | Reviewing goals and creating task plan |
| Working | Executing assigned tasks |
| Reviewing | Checking completed work |
| Paused | Manually paused by you |
Monitoring Agents
The dashboard shows real-time agent activity:
- Current task and progress
- Token usage and costs
- Recent outputs and decisions
- Communication between agents
Cost Control
Set per-agent budgets to control spending:
- Per-heartbeat limit — Maximum tokens per cycle
- Daily limit — Maximum tokens per day
- Monthly limit — Hard cap on monthly spending
When an agent hits its budget, it pauses and notifies you.