One line of code between your agent and a catastrophe.

Wrap the dangerous action. Ship the same day.

No proxy to deploy. No data path to reroute. You call the gate at the decision point; EMILIA answers allow, require signoff, or deny — decided by a policy engine with 26 machine-checked safety theorems behind it.

The HTTP call below is live. The SDK wrapper is the ergonomic version of the same thing. Exact signatures in the docs.

# Ask EMILIA whether this exact action may proceed — before it runs.
curl -s https://www.emiliaprotocol.ai/api/trust/gate \
  -H 'content-type: application/json' \
  -d '{
    "actor":   "agent_invoice_bot",
    "action":  "payment.release",
    "context": { "amount": 50000, "destination": "acct_9f12" }
  }'

# → { "decision": "allow_with_signoff",
#     "reason": "ai_agent_payment_action",
#     "signoff_required": true }

import { guard } from '@emilia-protocol/sdk';

// Wrap anything irreversible. One line.
const decision = await guard({
  actor:   agent.id,
  action:  'payment.release',
  context: { amount: 50_000, destination: invoice.account },
});

if (decision.deny) throw new Error('Blocked by policy');
if (decision.signoffRequired) {
  await decision.waitForHuman();   // blocks until a named human approves
}

// Proceeds only with a signed, verifiable Trust Receipt:
await bank.wire(invoice);          // decision.receipt is your audit evidence

Your agent is about to do something it can’t take back. One guard() call routes the exact action — actor, intent, parameters — to EMILIA before it touches the real world.

The formally verified policy engine returns one of three answers: allow (safe, proceed), allow-with-signoff (a named human must approve this exact action), or deny (a hard rule says no). No ambiguity, no silent pass.

Every decision emits a signed, Merkle-anchored Trust Receipt — who approved what, when, bound to the action hash. Auditor-grade evidence, verifiable offline with @emilia-protocol/verify. No EMILIA server required to check it.

Four things you never want an agent to do unsupervised.

Agent tries to wire $50K

An invoice-paying agent attempts a payment to a new account. Blocked. A named human signs off. Wire proceeds with a receipt.

Agent tries to drop prod

A coding agent runs a destructive migration on the production database. The action is gated before it executes.

Agent redirects a benefit

An agent changes the bank account on a benefits case. Blocked pending an accountable caseworker signoff — due process, proven.

Agent exfiltrates PII

An agent attempts to export a table of personal records to an external destination. The high-risk write is intercepted.