Is a payment-risk oracle the same as an x402 facilitator?

No. A facilitator verifies the payment signature and settles the payment on-chain — it is in the money path. A payment-risk oracle only returns a verdict on whether the payment is safe to make, before signing, and never touches funds.

How is a payment-risk oracle different from KYT or sanctions screening?

KYT/sanctions screening checks whether a wallet is on a sanctions list or tied to illicit flows. A payment-risk oracle additionally judges the counterparty's payment behavior (settlement and dispute history) and whether the quoted price is fair, and returns a single pre-signature verdict.

Updated 2026-06-30 · glossary

What is a payment-risk oracle?

A payment-risk oracle is a service an AI agent calls before authorizing a payment, to get an independent risk verdict on the counterparty and the amount. For the x402 protocol it runs in the pre-signature window — after the 402 Payment Required challenge, before the agent signs — and returns a GO / HOLD / STOP style verdict.

Why an agent needs one

With x402, an AI agent can pay for an API call, a dataset, or a service on its own — it receives a 402, signs a payment, and continues. That autonomy is the point, and also the exposure. Nothing in the protocol tells the agent whether the payee is trustworthy or whether the price is reasonable. An agent will sign whatever it is quoted. A payment-risk oracle is the independent check between the quote and the signature.

It answers a question the payment rails do not: should my agent trust this payee with this payment, at this price?

How it works

The oracle sits in the pre-signature window and takes the payment the agent is about to make — counterparty (payee), amount, asset, chain — and returns a verdict:

GO — proceed and sign.
HOLD — resolvable risk; escalate to a spending cap or a human.
STOP — do not sign.

The verdict is driven by evidence about the counterparty and the amount — typically the counterparty's settlement and dispute history (reputation) and whether the quoted amount is anomalous versus what that counterparty normally charges (price-anomaly). Crucially, a payment-risk oracle is advisory: it returns a verdict and is never in the settlement path. The agent decides.

What it is not

Not a facilitator. A facilitator verifies the signature and settles the payment on-chain. A payment-risk oracle only judges the payment, before it is signed, and never moves funds.
Not custody or escrow. It holds nothing.
Not KYC of the agent. It scores the payee so the buyer can decide — the reverse of most trust tooling, which scores the buyer so a seller can gate.

Related terms

An oracle's verdict often composes with two adjacent checks: endpoint readiness (is the payee's service set up correctly and within budget?) and KYT / sanctions screening (is the wallet sanctioned?). These are complementary layers, not substitutes — see the comparison of x402 payment-risk oracles.

Blackwall: a payment-risk oracle for x402

Blackwall is a payment-risk oracle. It returns GO / HOLD / STOP before an agent signs an x402 payment, from behavioral counterparty reputation (on-chain settlement history, with a dispute signal that activates as outcomes accrue) and price-anomaly detection — the quoted amount versus the counterparty's own per-payer median, with a STOP at ≥8×. Verdict only, never custody. You can get a verdict live:

curl -sS -X POST https://agent-egress-proxy.onrender.com/v1/forecast-payment \
  -H 'content-type: application/json' \
  -d '{"counterparty":"0x…","amount":"5.00","asset":"USDC","chain":"base"}'

Blackwall is live on Base (mainnet-configured). It is also an MCP server (forecast_payment, stdio). Honest status: pre-traffic; OFAC sanctions screening is live (best-effort, refreshed from the 0xB10C feed; not legal advice).