Loading…
Deterministic prepaid billing for humans in chat and systems in production.
Spend control that behaves deterministically under retries, streaming, and failures.
Reserve budget using max_tokens before execution.
Run the provider call if a compliant route exists.
Finalize usage and apply plan-based billing controls after completion.
Operational clarity across wallet, limits, and usage - by member, model, and time.
Balance, reserved, available - all evaluated deterministically.
Filter by member, model, time range, and request ID.
Require EU routes and ZDR retention mode per request, consistently across workspace and API.