The control plane for AI agent payments.

Wrap your agent's fetch once. Every x402 paywall becomes a policy decision before a cent moves.

import { createGuard } from '@reinconsole/sdk';

const guard = createGuard({ engineUrl, agentId });
const fetch = guard.wrap();

// vendor answers 402 → intent → policy check →
// signed decision. Deny blocks before payment exists.
const res = await fetch('https://api.vendor.example');

• budgets & tx caps
• allow / deny lists
• kill switch
• signed receipts

Price your routes once. Every payment into your API is screened, settled, and receipted before your handler runs.

import { createGate, gateMiddleware } from '@reinconsole/gate';

const gate = createGate({
  routes: [{ path: '/api/answer', price: '0.05' }],
  rails, payTo: '0xYourTreasury…',
  screen: { check: payerCheck(graph, { denyBelow: 40 }) },
});

app.use(gateMiddleware(gate));

• quote consistency
• replay protection
• velocity caps
• payer screening

Evidence in, scores out. Both sides of the wire feed one graph — and the graph feeds back into enforcement.

import { ReputationGraph, payerCheck } from '@reinconsole/graph';

const graph = new ReputationGraph()
  .observe(engine).observe(indexer).observe(gate);

// pushed scores make vendorReputationLt policies fire
await graph.syncVendors(engine.spend);

• never stored, always explainable
• unknown ≠ bad
• survives key rotation