Link Preview Metadata: Open Graph, Twitter card and favicon for any URL

Give it URLs and get back what a link unfurler (Slack, iMessage, X, LinkedIn, an AI agent) would show: title, description, preview image, site name and favicon, plus every Open Graph and Twitter card tag, canonical URL, language, author, published and modified dates, JSON-LD types, oEmbed endpoint, first H1 and a word count.

USD 1 per 1,000 URLs. Malformed URLs and pages that fail to load are free.

What people use it for

• Link previews in your app. Show rich cards for links your users paste, without running a headless browser or fighting with each site's quirks.
• AI agents and RAG. Give an LLM a clean summary of a URL (title, description, type, author, date, language, word count) before deciding whether to read the full page.
• Social card QA. Check that every page has og:title, og:description, og:image and a Twitter card before a launch.
• Content audits. Pull titles, descriptions, canonicals, authors, publish dates, H1s and word counts for a list of articles into one spreadsheet.
• Favicon and brand lookup. Get the best available icon for a domain, from <link rel="icon">, apple-touch-icon or the web app manifest.

Sample output

A real news article, trimmed:

{
  "input": "https://www.bbc.com/news/articles/cqx2z898p6p4o",
  "ok": true,
  "url": "https://www.bbc.com/news/articles/cqx2z898p6p4o",
  "statusCode": 200,
  "title": "Utility firms under fire over Ormskirk pavement repairs",
  "description": "Residents say streets have been left with temporary-looking repairs instead of matching stone flags.",
  "image": "https://ichef.bbci.co.uk/news/1024/branded_news/c2eb/live/c6c68a30-baa0-11f1-afa8-43b99a243ced.jpg",
  "type": "article",
  "canonicalUrl": "https://www.bbc.com/news/articles/cqx2z898p6p4o",
  "openGraph": {
    "og:title": "Utility firms under fire over Ormskirk pavement repairs",
    "og:type": "article",
    "og:image": "https://ichef.bbci.co.uk/news/1024/branded_news/c2eb/live/c6c68a30-baa0-11f1-afa8-43b99a243ced.jpg",
    "og:image:width": "1024",
    "og:image:height": "576"
  },
  "twitterCard": { "twitter:card": "summary_large_image", "twitter:title": "Utility firms under fire over Ormskirk pavement repairs" },
  "article": { "article:modified_time": "2026-09-27T19:20:10.231Z" },
  "language": "en-GB",
  "author": "Paul Faulkner",
  "publishedTime": "2026-09-27T19:20:10.231Z",
  "modifiedTime": "2026-09-27T19:20:10.231Z",
  "themeColor": "#ffffff",
  "jsonLdTypes": ["ReportageNewsArticle"],
  "oembedUrl": null,
  "h1": "Utility firms under fire over pavement repairs",
  "wordCount": 865,
  "favicon": "https://static.files.bbci.co.uk/bbcdotcom/web/20260922-091849-a8520cb5b4-web-3.22.0-3/android-chrome-512x512.png",
  "faviconSizes": "512x512",
  "faviconSource": "manifest"
}

A YouTube video adds the video tags and the oEmbed endpoint:

{
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "image": "https://i.ytimg.com/vi/dQw4w9WgXcQ/maxresdefault.jpg",
  "siteName": "YouTube",
  "type": "video.other",
  "openGraph": { "og:video:url": "https://www.youtube.com/embed/dQw4w9WgXcQ", "og:video:width": "1280", "og:video:height": "720" },
  "jsonLdTypes": ["VideoObject"],
  "publishedTime": "2009-10-24T23:57:33-07:00",
  "oembedUrl": "https://www.youtube.com/oembed?format=json&url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DdQw4w9WgXcQ",
  "favicon": "https://www.gstatic.com/youtube/img/branding/favicon/favicon_192x192_v2.png"
}

A page with nothing but a title (https://example.com) returns "title": "Example Domain", empty openGraph and twitterCard, and null for the image and description, so missing tags are easy to spot.

Field	What it is
title, description, image, siteName, type	The preview fields an unfurler shows: Open Graph first, then Twitter card, then the plain HTML <title> and meta description. image is made absolute.
htmlTitle, metaDescription	The plain HTML title and meta description, for SEO comparisons.
openGraph, twitterCard, article	Every og:*, twitter:* and article:* tag as found. Repeated tags (several og:image) become arrays.
canonicalUrl, language, author, themeColor	From <link rel="canonical">, <html lang>, meta author (or article:author or JSON-LD) and theme-color.
publishedTime, modifiedTime	From article:published_time / modified_time, og:updated_time or JSON-LD dates.
jsonLdTypes	Schema.org types in the page's JSON-LD, such as NewsArticle, Product, Recipe, Organization.
oembedUrl	The oEmbed endpoint the page advertises, if any.
favicon, faviconSizes, faviconSource, icons	The best icon (the largest declared, from link tags or the manifest; /favicon.ico only if it really exists), plus every icon found.
h1, wordCount	The first H1 and an estimate of the visible words (scripts and styles excluded).
url, statusCode, redirected, contentType	The final URL after redirects and its response.
ok, errorType, error	ok is false for pages that could not be loaded, with DNS, TLS, TIMEOUT, HTTP_ERROR and so on. Those rows are free.

How to use it

1. Paste URLs into URLs, one per line, or send a JSON array through the API.
2. Optional: set Fetch as to Chrome on Android to see the mobile page.
3. Click Start. Export CSV, JSON or Excel, or call the actor from your app through the Apify API and read the dataset.

For an app, call it synchronously through the API (run-sync-get-dataset-items) with a handful of URLs and cache the results.

Pricing

What	Price
URL previewed (a page or file that loaded with 2xx)	USD 0.001 (USD 1 per 1,000)
Malformed URL, DNS or TLS error, timeout, 4xx or 5xx	Free
Duplicate URLs in the list	Fetched and billed once

Set a maximum cost per run in the run options; the actor stops cleanly when it is reached.

Limits

• It reads the HTML the server sends, like Slack and Facebook do, and does not run JavaScript. Sites that add their tags with JavaScript only show what is in the raw HTML.
• Up to 5 MB of HTML per page. Image dimensions are not measured; use og:image:width and og:image:height when the site provides them.
• Some sites block data-center traffic. Use a proxy if you see 403 or 429 for sites that work in your browser.

Related actors

• Website Screenshot API: Full Page, Mobile, PDF: Screenshot any list of URLs as full-page or viewport PNG, JPEG, WebP or PDF, on desktop, laptop, tablet or mobile.
• Website Tech Stack Detector (Wappalyzer Alternative): Find the technology behind any website in bulk: CMS, ecommerce platform, analytics, frameworks, CDN, hosting, payment, email and...

FAQ

Why is the title different from what I see in the browser tab? Link previews use og:title first. The browser tab uses <title>, which is in htmlTitle.

Does it fetch the oEmbed data? It returns the endpoint URL. Call it yourself if you need the embed HTML.

Is it legal? It requests public URLs that you supply, the same way a browser or a chat app's link preview does, and reads the page's own metadata. It does not log in and collects no personal data beyond what the page publishes about itself (for example the article's author line).

Something missing or wrong? Open an issue with the URL.