Trust Check

Give it a business URL, get back an explainable legitimacy score built from public signals. For an agent about to transact with a counterparty it has never seen, "is this real?" is the question worth five cents before the money moves.

• Six signals, weighted by how expensive each is to fake: domain age (RDAP, from the registry of record), the live TLS certificate, contact surface, policy pages, postal address, and review footprint
• Every signal carries its own evidence string — the registration date, the certificate issuer, the path of the privacy page — so a score can be audited rather than trusted
• Deterministic: no language model participates in the scoring, and the entire model (weights, cutoffs, formula) is published at GET /api/pricing and the MCP get_scoring_model tool, so any caller can recompute a score it was handed
• A signal that could not be observed leaves the DENOMINATOR rather than scoring zero, and the response reports how many of the six resolved. Too few, and the band is "insufficient" with score null — no number is invented
• 5 credits per check; a check where not one signal resolved is refunded in full
• KYB-lite, not a verdict: it reports public signals, does not certify a business, and a low score usually means NEW, not dishonest
• Agent-native: trs_live_ REST keys, an MCP server (works with the Ounie AI Team via ?api_key=), keyless x402 pay-per-call in USDC, and the fleet master key

{
  "mcpServers": {
    "ounie": {
      "url": "https://mcp.ounie.com"
    }
  }
}

{
  "mcpServers": {
    "trustcheck": {
      "url": "https://trustcheck.ounie.com/api/mcp"
    }
  }
}