Link preview and Open Graph metadata API
Get title, description, canonical URL, Open Graph and Twitter card fields, favicon, language, feeds and JSON-LD types for any public URL. Pay per page.
Chat apps, dashboards and agents all need the same small thing: what is behind this link. The call reads one or many public URLs and returns the title, description, canonical URL, Open Graph and Twitter card fields, favicon, language, hreflang alternates, feed links and the JSON-LD types found on the page.
It respects robots.txt, refuses private addresses and does not run JavaScript, so it is fast and predictable. Each page is one result; pages that fail are listed with the reason and not charged.
The calls
Each call returns 402 with the price until a payment is attached; an x402 client pays and retries automatically. Calls that fail are not charged. See Get started and the guide for bots and agents for code.
url-metadata $0.003 per call
Link-preview data for one or many public URLs: title, description, canonical URL, Open Graph and Twitter card fields, favicon, language, hreflang alternates, feed links and JSON-LD types.
curl -i "https://api.ambolt.dev/v1/url-metadata?url=https%3A%2F%2Fwww.sitemaps.org%2F"Example response
{
"results": [
{
"url": "https://www.sitemaps.org/",
"ok": true,
"finalUrl": "https://www.sitemaps.org/",
"status": 200,
"contentType": "text/html",
"redirected": false,
"title": "sitemaps.org - Home",
"description": "The Sitemaps protocol enables webmasters to information earch engine about pages on their site that are available for crawling.",
"h1": "sitemaps.org",
"language": "en",
"canonical": null,
"robotsMeta": null,
"openGraph": {},
"twitter": {},
"favicon": "https://www.sitemaps.org/favicon.ico",
"hreflang": [],
"feeds": [],
"jsonLdTypes": [],
"themeColor": null,
"generator": null
}
],
"succeeded": 1,
"failed": 0,
"billableResults": 1,
"fetchedAt": "2026-10-02T23:25:07.457Z",
"note": "Metadata is the publisher's content; check its licence before reuse."
}Questions
Does it render JavaScript?
No. It reads the HTML that the server returns, which is what link previews use.
Can I send many URLs?
Yes. Pass the urls parameter with a list; each page is one result.
Does it respect robots.txt?
Yes, and it refuses private and local addresses.
Use it from an AI agent
Every call is also an MCP tool: add https://api.ambolt.dev/mcp to your agent's server list.