Cryptographic attestation for Claude agents.

Native MCP integration. Attest any Claude tool call or decision with verifiable human oversight, anchored on-chain.

How it wires up

1. Your Claude agent connects to SanctifAI Trust via MCP at https://trust.sanctifai.com/mcp.
2. When the agent reaches a decision that needs human oversight, it calls the request_verification tool.
3. A qualified human attests the decision via WebAuthn (Touch ID, security key, etc.).
4. The attestation is anchored on Base via EAS. The agent receives a verifiable certificate URL.

Connecting Claude (Claude Code MCP)

claude mcp add sanctifai-trust https://trust.sanctifai.com/mcp \
  --header "Authorization: Bearer sk_live_..."

In an Anthropic SDK call

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic();

const message = await client.messages.create({
  model: "claude-opus-4-7",
  max_tokens: 1024,
  mcp_servers: [{
    type: "url",
    url: "https://trust.sanctifai.com/mcp",
    authorization_token: process.env.SANCTIFAI_API_KEY,
    name: "sanctifai-trust"
  }],
  messages: [{ role: "user", content: "Process this loan application." }]
});