Model Context Protocol for autonomous payments
Enable AI agents to make and receive payments autonomously using the Model Context Protocol (MCP). Built for the agentic economy with budget controls and audit trails.
AI agents can pay for resources, APIs, and services without human intervention.
Set spending limits, approved merchants, and transaction rules for your agents.
Every agent transaction is logged with context about why the payment was made.
Manage payments across fleets of agents with hierarchical permissions.
The @t402/mcp package provides MCP tools for payment operations. Agents can check balances, make payments, and verify transactions.
Standard MCP tools: t402_pay, t402_balance, t402_verify, t402_history. Compatible with Claude, GPT, and other MCP-enabled models.
Define rules in JSON: max per-transaction, daily limits, approved recipients, required confirmations for large amounts.
Agents operate with derived keys or smart contract wallets. Parent accounts maintain ultimate control.
Get started with our quickstart guide and have ai agent payments (mcp) working in minutes.
Get Startedimport { createMCPServer } from "@t402/mcp";
const server = createMCPServer({
wallet: agentWallet,
policies: {
maxPerTransaction: "10.00",
dailyLimit: "100.00",
approvedRecipients: ["0x...", "0x..."],
},
networks: ["eip155:8453", "eip155:42161"],
});
// Agent can now use t402_pay tool
// Claude: "Pay $5 USDT0 to the API provider"
// → t402_pay({ to: "0x...", amount: "5.00", network: "eip155:8453" })Check out the documentation for detailed implementation guides and API references.