Proof of Human for OpenAI Agents.

Trust is a participation seal, not identity and not correctness. Assistants and the Agents API have no browser WebAuthn — use the hosted chat bridge.

How it wires up

1. Your agent reaches a decision that needs a human to show up and seal a payload.
2. Call POST https://bridge.trust.sanctifai.com/api/v1/attestations. Do not send a Trust API key from the agent.
3. A human opens approve_url in Chrome and confirms with a passkey.
4. Poll until status=completed and keep certificate_url.

As an OpenAI tool

const tools = [{
  type: "function",
  function: {
    name: "request_human_seal",
    description: "Mint a SanctifAI Trust participation seal via the chat bridge.",
    parameters: {
      type: "object",
      properties: {
        task_subtype: { type: "string" },
        taskData: { type: "object" },
        resultData: { type: "object" }
      },
      required: ["task_subtype", "taskData", "resultData"]
    }
  }
}];

async function requestHumanSeal(args) {
  const res = await fetch("https://bridge.trust.sanctifai.com/api/v1/attestations", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({
      task_type: "GEN",
      domain: "GEN",
      requesting_agent: "OpenAI agent",
      ...args,
    }),
  });
  return res.json(); // { attestation_id, approve_url, status: "pending" }
}