Every Action Logged: PaperclipAI's Activity Trail and Cost Tracking
PaperclipAI maintains a complete activity log of every mutation in the system. Agent created, task status changed, approval granted, heartbeat completed — every event is recorded with timestamps. Combined with the cost tracking (provider, model, tokens, cents per heartbeat), you get full operational transparency.
What the activity log captures
- Agent lifecycle events: creation, state changes (active, paused, error, terminated)
- Task workflow: status transitions through backlog → todo → in_progress → in_review → done
- Approval decisions: pending, approved, rejected, revision requested
- Heartbeat results: execution output, duration, and any errors
- Cost data: provider, model, input/output tokens, cost in cents per heartbeat
- Board operator actions: overrides, manual agent management, direct task assignments
Cost reporting endpoints
| Endpoint | What It Shows |
|---|---|
| /api/companies/{id}/costs/summary | Total company spend this month |
| /api/companies/{id}/costs/by-agent | Per-agent cost breakdown |
| /api/companies/{id}/costs/by-project | Cost attributed to each project/goal |
The dashboard surfaces this data visually: current-month spend vs. budget for the company and each agent. Anomalies — like a sudden spike in one agent's costs — become immediately visible.
Keep reading
When Agents Get It Wrong: PaperclipAI's Built-In Error Handling and Oversight
Agents fail heartbeats, produce bad output, or exceed budgets. PaperclipAI catches this with agent states, board overrides, approval gates, and automatic budget pausing.
Board Operator Workload: How Much Time Does PaperclipAI Actually Require?
Active setup phase: 1-2 hours/day. Steady state: 15-30 minutes/day reviewing dashboards and approvals. The oversight load decreases as your agents' instructions mature.
Want to talk about what AI can do for your business?
30 minutes, no sales pitch. Just an honest conversation about where AI can move the needle for you.
Book a Free Call