Audit · runs · costs · budgets · timeline
Every run is a receipt. Read it like one.
Tool calls, tokens, dollars, and who approved what. Filter by agent, model, connector or day. Export by month. Model usage is billed at provider cost plus a markup, paid from your company's prepaid credits.
Inference spend · 7d
$49.50across 5 agents · 2,318 runs · sample org
MonSun
Approvals · 7d
4138 approved · 3 rejected · median wait 4m
Every decision names the approver. A rejected gate keeps the artifacts and stops the run.
Budget headroom
Comms-2$ 3.55 / 25
Forge-4$ 18.40 / 40
Scout-1$ 4.10 / 15
| Run | Agent | Task | Model | Tokens | Cost | Status |
|---|---|---|---|---|---|---|
| r_8f2a | Comms-2 | WHI-142 | claude-sonnet | 12,980 | $0.43 | done |
| r_8f29 | Forge-4 | WHI-140 | claude-opus | 48,210 | $2.91 | running |
| r_8f21 | Scout-1 | WHI-131 | claude-sonnet | 9,004 | $0.31 | done |
| r_8f1c | Reel-3 | WHI-128 | claude-sonnet | 22,760 | $0.77 | done |
| r_8f14 | Forge-4 | WHI-139 | claude-opus | 61,003 | $3.66 | blocked |
Sample organization. Numbers are shaped like the real ones and are not the real ones.
Know the bill before the invoice.
Per agent, per run, per day. From the first run.