Open Graph & Link Preview Extractor

Extract the link-preview metadata a page publishes for itself - what Slack, Twitter/X and LinkedIn read when you paste a URL. Returns title, description, Open Graph & Twitter-card tags, favicon, JSON-LD, images. Pure code, no proxy/AI. Reads PUBLIC meta only; does NOT scrape private social accounts.

Pricing

$1.00 / 1,000 results

Rating

0.0

(0)

Developer

Ahmed Moussa

Actor stats

Bookmarked

Total users

Monthly active users

23 days ago

Last modified

What it does

Given a public URL (or a small bounded batch), it fetches the page once and deterministically extracts:

title, description, canonical_url, site_name, lang, theme_color, favicon
og — Open Graph block: {title, description, image, type, url}
twitter — Twitter Card block: {card, title, description, image, creator, site}
json_ld[] — schema.org JSON-LD objects the page publishes
images[] — representative image URLs (og/twitter image first, then page <img> sources), absolutized & deduped
method (which layer produced the data) and a code-owned parse_confidence

No proxy, no headless browser, no LLM, no paid API. $0 idle, $0 uncovered beyond Apify compute.

Input

Field	Type	Description
`url`	string	A single public URL to build a preview for (include `https://`).
`urls`	array	Optional bounded list of additional URLs (capped at 50/run).

{
  "url": "https://github.com/apify/apify-sdk-python",
  "urls": [
    "https://www.theverge.com/",
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  ]
}

Output

One record per URL in the dataset:

{
  "url": "https://github.com/apify/apify-sdk-python",
  "status": "completed",
  "title": "GitHub - apify/apify-sdk-python: The Apify SDK for Python ...",
  "description": "The Apify SDK for Python is the official library for creating Apify Actors ...",
  "canonical_url": "https://github.com/apify/apify-sdk-python",
  "site_name": "GitHub",
  "og": {
    "title": "GitHub - apify/apify-sdk-python: The Apify SDK for Python ...",
    "description": "The Apify SDK for Python ...",
    "image": "https://opengraph.githubassets.com/.../apify/apify-sdk-python",
    "type": "object",
    "url": "https://github.com/apify/apify-sdk-python"
  },
  "twitter": {
    "card": "summary_large_image",
    "title": "GitHub - apify/apify-sdk-python ...",
    "description": "The Apify SDK for Python ...",
    "image": "https://opengraph.githubassets.com/.../apify/apify-sdk-python",
    "creator": null,
    "site": "@github"
  },
  "favicon": "https://github.com/favicon.ico",
  "json_ld": [ { "@type": "WebSite", "...": "..." } ],
  "theme_color": "#1e2327",
  "lang": "en",
  "images": [ "https://opengraph.githubassets.com/.../apify/apify-sdk-python" ],
  "parse_confidence": "high",
  "method": "opengraph",
  "extracted_at": "2026-06-24T10:00:00+00:00",
  "error": null
}

Use cases

Link previews — render unfurled share cards for any URL in your own app/chat/CMS.
SEO & social audit — check that pages publish valid og:title/og:image/twitter:card before they go live.
Card validation — catch missing or oversized preview images and broken canonical/site-name tags in bulk.

How it works

SSRF-guarded fetch — the target host is resolved and rejected if it points at a private/loopback/reserved IP; a domain blocklist refuses login-walled/ToS-hostile hosts; redirects are re-validated per hop. A single streamed GET with hard size (2 MB) and time caps fetches the page.
Deterministic parse — pure-code extraction of <meta> (Open Graph + Twitter + standard), <title>, <link rel=canonical/icon>, <html lang>, JSON-LD and <img> sources. First occurrence wins (matches how preview platforms read tags). Image URLs are absolutized against the final URL.
Honest confidence — parse_confidence is computed in code from how many Open Graph / Twitter signals were actually present (high/medium/low/none); method reports the richest layer used.

Limitations (honest)

Reads only what the page publishes server-side in its HTML. Pages that render Open Graph tags client-side via JavaScript (no SSR) will return little — there is no headless browser by design (cost-safety).
No private social data. Followers, posts, engagement, DMs, private profiles, paginated feeds: out of scope and not possible without paid proxies/ToS violations.
Login-walled / ToS-restricted hosts (e.g. linkedin.com, facebook.com, instagram.com) are intentionally blocked and return status: "blocked".
One page per record — the actor never follows links or paginates.
parse_confidence/method are heuristics over which tags were found, not a guarantee of correctness of the publisher's own values.

Link Preview & URL Metadata API — Open Graph Extractor

f0rty7even/link-preview

Free API: turn any URL into clean link-preview metadata — title, description, Open Graph & Twitter card, favicon, canonical, and language. Powers link unfurling and preview cards. No login.