Regent Protocol Start building
Platform · Mandates

The rules money cannot cross.

A mandate defines what an agent may do with money: amount, destination, time, conditions, and approvals — enforced before execution, not reviewed after.

The failure it prevents

Post-hoc controls find the bad payment in tomorrow's report. A mandate refuses it today: the transaction outside the rules is denied before any money moves, and the agent never holds the payment credentials itself.

Try the logic

Mandate simulator.

Adjust the rules and the action — watch the decision change. Runs entirely in your browser; illustrative only, nothing is sent anywhere.

Per-transaction limit
Allowed destination
Approval required above
Action: amount
Action: destination
NEEDS APPROVAL
Illustrative workflow — a local demo of mandate logic, not a live production integration. Real fields include daily/monthly ceilings, per-entity sub-budgets, and relational caps (a refund may not exceed its original charge).
Mechanics

Inputs, decision, outputs.

Inputs

The mandate's limits (per-transaction, daily, monthly, per-entity, relational) and the action: amount, currency, optional entity and reference.

Decision

Authorize-before-execution: within limits → an authorization with a short-lived signed proof (JWT). Outside → a coded refusal (LIMIT_EXCEEDED names the ceiling) and no movement. Owner approval can be required above thresholds.

Operator visibility

Every authorization and refusal lands in the audit trail and the Control Panel. Usage counters reconcile via settle when the real charge lands.