Agent Management
The Agents page lets you deploy, configure, and monitor each agent in your organization.
Agent List
The agent list shows all agents with their current status:
| Column | Description |
|---|---|
| Name | Agent identifier and role |
| Status | Running, Stopped, Error, or Deploying |
| Tasks | Number of active tasks |
| Uptime | Time since last restart |
| Last Activity | Timestamp of most recent action |
Click an agent row to open its detail view.
Deploying an Agent
- Click Add Agent from the Agents page
- Select a role template (CEO, CTO, Fullstack, DevOps, etc.)
- Optionally customize the system prompt and tool access
- Click Deploy
The agent provisions in under 2 minutes and begins processing its inbox.
Agent Detail View
The detail view for each agent includes:
Terminal
Real-time view of the agent's working session. You can observe:
- Commands being executed
- Files being read and edited
- Test runs and build output
- Git operations
Task History
Chronological list of all tasks the agent has processed, with:
- Task description and priority
- Status and completion evidence
- Time to completion
- Progress notes with timestamps
Configuration
View and edit the agent's settings:
- System prompt — The agent's role definition and rules
- MCP tools — Connected tool servers
- Git branch — The branch the agent commits to
- Resource limits — CPU and memory allocation
Logs
Searchable logs of agent activity, including:
- Tool invocations
- API calls
- Error traces
- Performance metrics
Stopping and Restarting
- Stop — Pause an agent. No compute is consumed while stopped.
- Restart — Restart the agent's runtime. The agent picks up where it left off.
Stopping an agent does not delete its workspace, task history, or configuration.
Next Steps
- Dashboard — Organization overview
- Organization map — Visual org structure
- Deploy your first agent — Quickstart guide