USEFUL APIS FOR AI AGENTS
Reader
Clean article text, Markdown and tables.
Extract a readable article from supplied HTML, including title, author, language, metadata, clean Markdown, links and tables. Uses Mozilla Readability with a fallback for short pages. No remote fetches or script execution; URLs resolve relative links.
Make this part of your workflow.
One fixed example is free. Custom work uses a clear input contract and releases the result after payment settles.
One copyable request.
POST /v1/articles/extract
{
"url": "https://example.com/notes/launch",
"html": "<!doctype html><html lang=\"en\"><head><title>Launch notes</title><meta name=\"description\" content=\"Release handoff notes\"></head><body><nav>Home | Pricing | Contact</nav><article><h1>Launch notes</h1><p>The release is ready for review. The deployment checklist includes payment verification, public documentation, and customer support handoff.</p><p>Each service has a clear contract and a working example. The team will monitor errors, latency, and successful settlement receipts throughout the first week.</p><h2>Milestones</h2><table><tr><th>Task</th><th>Status</th></tr><tr><td>Payment verification</td><td>Complete</td></tr></table><p>Read the <a href=\"/checklist\">full checklist</a> before approving the release.</p></article><footer>Copyright Example</footer></body></html>"
}Send the JSON to the operation above. An x402-capable client receives the payment challenge, authorizes exactly 0.02 USDC, and retries with PAYMENT-SIGNATURE.
Keep the result.
Markdown + text + tables + links
Files arrive in a JSON artifact envelope with base64 bytes, filename, MIME type, size and SHA-256. Decode the bytes to save the file. Diagnostic results arrive as structured JSON.
The same proof and normalized input retrieves the exact paid result for 30 minutes across restarts. A proof cannot unlock another request. Never create another authorization automatically after ambiguous delivery.