Spend

Cost per agent, run, customer and model — derived from the record itself.

Sample data⏻ halt

Total

$0.1554

7 runs

Clean runs

5/7

no block, no human needed

Cost per clean run

$0.0311

unit economics

Unpriced calls

upstream returned no cost

2 model call(s) could not be priced by the upstream. They are recorded as unattributed rather than as $0 — the total above is a floor, not a complete figure.

Budgets

none set

Ceilings per agent or fleet-wide, enforced at decision time from the chain itself — an agent cannot understate what it has already spent.

Set a budget

No ceiling is set, so none is enforced. An agent's own budget wins over the fleet default, and spend is summed from the chain at decision time — so an agent cannot understate what it has already spent.

By agent

share of spend

appointment-setter$0.1190

2 runs

loan-underwriter$0.0364

1 run

payments-bot$0.0000

1 run

release-agent$0.0000

1 run2 blocked

infra-agent$0.0000

0 runs

Cost breakdown

where the money goes

Input$0.0206

Output$0.0338

Reasoning$0.0000

Cache read$0.0000

Cache write$0.0000

Tool usage$0.0700

Reasoning is a subset of output; cache costs are subsets of input. They are shown separately, never summed into the total.

By customer

margin per account

cust-77$0.0364

cust-208$0.0000

cust-209$0.0000

By model

claude-sonnet-4-5

2 calls

$0.0364

acme-internal-reranker-v2

2 calls

$0.0000