Guard decisionREADY
Tap Run demo to show the selected source, decision, and audit proof.
AI payment safety layer
AgentPay Guard
AI payment guardrail that allows, reviews, or blocks agent spend with visible audit proof.
Deterministic policyVisible evidenceCitePay in the loop
How it fits in the stack
AI AgentAgentPay Guardx402 / Circle GatewayPaid API / Service
Proposed0 USDC
Allowed0 USDC
Review0
Blocked0
Main demo narrative
CitePay flow with Guard in the loop
Run one query, inspect the paid-source candidates, and show how every spend request is allowed, reviewed, or blocked with audit evidence.
0 selected0 approved
01
Agent request
What the agent is trying to buy
Lepton/CitePay demo preset0.24 USDC budget • local demo preset
02
Selected paid sources
Which candidates make it into the spend set
Proposed: 0 USDCAllowed: 0 USDC
No selected sources yet.
Run the preset to show how the agent picks candidate sources before any spend is evaluated.
Skipped sourcesOptional details
No skipped sources yet.
03
Guard decisions
What the control layer allows, reviews, or blocks
No decisions yet.
After the flow runs, each selected source gets an explicit decision, risk score, policy match, and audit ID.
Trust + evidence
Decision proof
Latest decisionREADY
Tap Run demo to show the selected source, decision, and audit proof.
Full audit logExpand for machine-readable history
Recent audit log
Machine-readable history of recent decisions and spend intents.
| Time | Decision | Agent | Amount | Recipient | Audit ID |
|---|---|---|---|---|---|
| No audit records yet. | |||||
Policy test casesExpand validator mode
Validator mode
Policy test cases
Editable payment intents for deterministic ALLOW / REVIEW / BLOCK proofs.
Scenario
Decision
No validator output yet.
Select a test case and evaluate it to show a deterministic policy result.