Counterparty Trust Check: ALLOW/REVIEW/DENY before you pay
Pricing
from $10.00 / 1,000 trust verdicts
Counterparty Trust Check: ALLOW/REVIEW/DENY before you pay
Before an agent pays an invoice, accepts new bank details, onboards a vendor or emails a new counterparty: one call returns ALLOW / REVIEW / DENY, a 0-100 score, reasons, next actions, and what changed since your last check.
Pricing
from $10.00 / 1,000 trust verdicts
Rating
0.0
(0)
Developer
Probelane
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
2 days ago
Last modified
Categories
Share
Counterparty Trust Check
One call, one verdict: should your agent trust this domain?
Use it before an AI agent (or a workflow) pays an invoice, sends money or data, signs up, accepts a webhook, or emails a new counterparty. Input a domain, URL or email address; get back ALLOW, REVIEW or DENY with a 0–100 score, plain-language reasons and the raw public evidence.
When an agent should pick this tool
- A payment request, invoice or wallet address arrives from an unfamiliar domain.
- A new vendor, supplier or API provider must be onboarded.
- An inbound email asks for money, credentials or a change of bank details (check the sender's domain).
- A link or checkout page must be judged before entering details.
Input
{"domains": ["stripe.com", "billing@new-vendor.xyz"]}
Output (one dataset item per domain)
{"domain":"nsfas.org.za","verdict":"ALLOW","score":80,"flags":["no DMARC: email from this domain can be spoofed"],"evidence":{"resolves":true,"mx":["nsfas-org-za.mail.protection.outlook.com"],"spf":"v=spf1 ...","dmarc":{"record":null},"domain_age_days":null,"tls":{"valid":true}},"checked_utc":"2026-10-07T14:40:00+00:00"}
Built for repeat use in agent workflows
Every verdict carries:
recheck_after_utc— when this counterparty should be checked again (24 h for DENY/REVIEW, 72 h for young domains, 7 days otherwise).changed_since_last_check+changes— compared with your own previous check of the same domain (DMARC policy, MX, SPF, TLS issuer, score, verdict). A counterparty whose mail setup changes right before a payment request is the classic invoice-fraud pattern.next_actions— machine tokens such asdo_not_pay_or_send_data,require_out_of_band_confirmation_before_payment,verify_payment_or_bank_detail_requests_by_second_channel.history— your last 20 verdicts for that domain.
History is kept in a named key-value store in your own Apify account (counterparty-trust-history), readable and exportable any time. Turn it off with "rememberHistory": false.
Recommended pattern: call before every payment, onboarding or bank-detail change; schedule a re-check of your active vendor list at recheck_after_utc.
Signals
Domain existence (NXDOMAIN → DENY), domain age via RDAP (<30 days heavily penalised), valid TLS certificate, MX present, SPF, DMARC policy (missing / p=none means email from the domain can be spoofed), registry hold status.
Pricing
Pay per event: $0.01 per verdict (one event per unique domain). No platform usage charges. Typical latency 1–3 s per domain. Payable by AI agents with USDC via x402 where eligible.
Limits
Public signals only; a verdict is a risk signal, not a guarantee. Some ccTLDs (e.g. .co.za) do not publish registration dates over RDAP; this is flagged, not hidden.