Skip to main content
GET
Net Judge spend, broken down by service type, plus saved-run counts by kind. Refunds are already netted off: a failed run appears as a charge and a matching refund, and contributes nothing to net_spend_usd.

Request

Headers

string
required
Bearer token — your standard 60db API key (sk_…), or a user JWT. There is no separate Judge credential.

Query parameters

string
default:"current_month"
current_month, last_30_days or all_time. An unrecognised value falls back to current_month.

Response

object
net_spend_usd, operations (billed calls) and refunds.
object
Keyed by JUDGE_EVALUATE, JUDGE_EXTRACT and their _REFUND twins. gross_units is input tokens.
object
Saved-run counts per kind, with question totals and average lowest-confidence. A run sent with save: false is billed but never stored, so these can trail by_service.

Example