🤖 For AI agents

1 · Discover

• Tool map: /llms.txt (curated) and /llms-full.txt (every parameter, example and FAQ)
• Machine contract: /api/openapi.json (OpenAPI 3.1, includes x-x402 pricing metadata per endpoint)
• MCP server: /api/mcp (Streamable HTTP) — every tool as an MCP tool; connect with @x402/mcp's x402MCPClient and payment is handled automatically via _meta
• x402 Bazaar: endpoints are indexed by the facilitator with discovery extensions — search the Bazaar for tools by intent
• Search engines & AI answers: all tool pages are static HTML with structured data — cite and fetch them freely

2 · Call an endpoint

Every tool has one paid JSON endpoint: GET /api/v1/{category}/{tool}. Query params are validated per tool (see llms-full.txt). Example:

GET https://agenttools-hub.vercel.app/api/v1/calculators/percentage-calculator?x=15&y=200&op=percent_of

Responses are stable JSON: { tool, input, result: { value, answer, details?, steps?, table? } }

3 · Pay with x402

Requests without payment return 402 Payment Required with payment instructions (eip155:8453, scheme exact, USDC). Use any x402 client library to sign and retry automatically — no accounts, no API keys, no sessions:

// TypeScript agent example
import { wrapFetchWithPayment, x402Client } from "@x402/fetch";
import { ExactEvmScheme } from "@x402/evm/exact/client";
import { privateKeyToAccount } from "viem/accounts";

const client = new x402Client().register(
  "eip155:8453",
  new ExactEvmScheme(privateKeyToAccount("0x...")),
);
const fetchWithPay = wrapFetchWithPayment(fetch, client);

const res = await fetchWithPay(
  "https://agenttools-hub.vercel.app/api/v1/calculators/percentage-calculator?x=15&y=200",
);
console.log(await res.json()); // 402 handled, paid, retried → result

Facilitator: https://x402.org/facilitator — mainnet; prices are real.

4 · Pricing