AI 服務暫時不可用，以下為來源正文，待恢復後補全翻譯。
Departures · Pay-per-call APIs for AI agents A marketplace of tools, priced by the call. No accounts, no subscriptions, no API keys. Every tool below — ours and everyone else's — is paid for in USDC via x402: request, get a 402, sign a payment, get your data. That's the whole integration. /extract data-transform $0.005 LIVE /scrape web-scraping $0.01 LIVE /status dev-tools $0.001 LIVE /robots-check dev-tools $0.001 LIVE Browse the full directory → List your tool — $2/30 days Why agent-tools.net For agents & developers Integrate in one request No signup flow to script around. Your agent hits an endpoint, reads the price off the 402 response, signs, and retries. The whole loop fits in a single function. For tool builders Distribution without a billing system List your own agent-facing tool for $2/30 days, paid the same way you sell calls — no Stripe, no accounts, no invoices. Show up in the directory other agents actually search. How a call actually works 01 Request normally. GET /extract?url=... — no header, no key, nothing to configure up front. 02 Get a 402. The PAYMENT-REQUIRED header carries the exact price, network (Base), asset (USDC), and address — machine-readable. 03 Sign and retry. Your agent signs an EIP-3009 authorization for that amount and resends with X-PAYMENT. Verified, settled, and answered in one round trip. Our own departures Four endpoints we run ourselves — also listed in the directory below, same as anyone else's. GET /extract · $0.005 /extract Fetch a URL server-side and return clean {title, text, links} extracted from its HTML — skip the HTML parsing. curl "https://agent-tools.net/extract?url=https://example.com" \ -H "X-PAYMENT: " GET /scrape · $0.01 /scrape Structured scraping with simple selectors (tag, tag.class, tag#id) — repeat &field= for multiple columns, zipped into rows by document order. curl "https://agent-tools.net/scrape?url=https://example.com&field=title=h2.product-title&field=price=span.price" \ -H "X-PAYMENT: " GET /status · $0.001 /status Is this URL up? Status code, final redirect target, response time, content-type — without downloading the page body. curl "https://agent-tools.net/status?url=https://example.com" \ -H "X-PAYMENT: " GET /robots-check · $0.001 /robots-check Will this URL let you fetch it? Checks robots.txt for a given (or your own) user-agent before you spend a call finding out the hard way. curl "https://agent-tools.net/robots-check?url=https://example.com" \ -H "X-PAYMENT: "