Pricing

from $2.40 / 1,000 preview returneds

Try for free

Go to Apify Store

Link Preview API - Open Graph, Twitter Card & oEmbed Metadata

Try for free

For chat apps, CMS editors and AI agents that unfurl links: send URLs, get title, description, image, site name, favicon, canonical URL and language from Open Graph, Twitter Card, oEmbed and the HTML head. About a second per URL. Respects robots.txt; URLs that cannot be previewed are free.

Pricing

from $2.40 / 1,000 preview returneds

Rating

0.0

(0)

Developer

NeverEmpty

Actor stats

Bookmarked

Total users

Monthly active users

3 days ago

Last modified

Input

Field	Type	Default	What it does
`urls`	array of strings	(empty: the example `https://github.com/apify/crawlee` is used)	Page URLs, 1 to 1,000 per run. A URL without `http://` or `https://` is read as `https://`. The same URL given twice (also when only the `#fragment` differs) is read and charged once.
`acceptLanguage`	string	(not sent)	Optional `Accept-Language` header, for sites that show a different language version by language, for example `en`, `ja-JP` or `de-DE,de;q=0.9`. Empty means the version the site shows by default.
`includeOembed`	boolean	`true`	When the page links to its own oEmbed JSON endpoint in its `<head>` (on 2026-09-24: YouTube and Spotify among the 23 pages read), also read it and add the `oembed` object.
`onlyChanges`	boolean	`false`	Monitor mode: return only URLs whose title, description, image, site name, canonical URL or type changed since the last run with the same `watchName`, plus URLs new to the watch.
`watchName`	string	(none)	Name of the remembered state (letters, digits, `.`, `-`, `_`; up to 40). Setting it fills `changeType` and the previous-value columns. With monitor mode on and no name, `default` is used.
`resetMonitoringState`	boolean	`false`	Forget what this watch remembered, so every URL is returned as a first check.
`maxConcurrency`	integer	`4`	URLs read in parallel, 1 to 8.
`requestTimeoutSecs`	integer	`20`	Time limit for one request, 5 to 60 seconds. HTTP 429, 500, 502, 503 and 504 are asked again up to two more times; a site that does not answer in time is asked once more.

{
  "urls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ", "github.com/apify/crawlee"]
}

Output

One row per URL. A real row from 2026-09-24 (long text shortened here):

{
  "status": "ok",
  "position": 1,
  "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "finalUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "httpStatus": 200,
  "contentType": "text/html; charset=utf-8",
  "redirects": [],
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "description": "The official video for “Never Gonna Give You Up” by Rick Astley. ...",
  "image": "https://i.ytimg.com/vi/dQw4w9WgXcQ/maxresdefault.jpg",
  "imageWidth": 1280,
  "imageHeight": 720,
  "imageAlt": null,
  "siteName": "YouTube",
  "favicon": "https://www.youtube.com/s/desktop/02b72088/img/favicon_32x32.png",
  "appleTouchIcon": null,
  "canonicalUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "ogUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "language": "en",
  "languageSource": "html",
  "type": "video.other",
  "twitterCard": "summary_large_image",
  "twitterSite": "@youtube",
  "author": "Rick Astley",
  "publishedTime": null,
  "modifiedTime": null,
  "themeColor": "rgba(255, 255, 255, 0.98)",
  "video": "https://www.youtube.com/embed/dQw4w9WgXcQ",
  "oembed": {
    "type": "video",
    "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
    "authorName": "Rick Astley",
    "authorUrl": "https://www.youtube.com/@RickAstleyYT",
    "providerName": "YouTube",
    "providerUrl": "https://www.youtube.com/",
    "thumbnailUrl": "https://i.ytimg.com/vi/dQw4w9WgXcQ/hqdefault.jpg",
    "thumbnailWidth": 480,
    "thumbnailHeight": 360,
    "html": "<iframe width=\"200\" height=\"113\" src=\"https://www....",
    "width": 200,
    "height": 113
  },
  "oembedEndpoint": "https://www.youtube.com/oembed?format=json&url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DdQw4w9WgXcQ",
  "jsonLdTypes": ["VideoObject"],
  "robotsMeta": null,
  "sources": ["openGraph", "oembed"],
  "openGraph": { "title": "...", "type": "video.other", "...": "..." },
  "twitter": { "card": "summary_large_image", "site": "@youtube", "...": "..." },
  "changeType": null,
  "changedFields": null,
  "previousTitle": null,
  "previousDescription": null,
  "previousImage": null,
  "previousCheckedAt": null,
  "charset": "utf-8",
  "headBytesRead": 716367,
  "headComplete": true,
  "watchName": null,
  "note": null,
  "fetchedAt": "2026-09-24T13:07:11.920Z"
}

Where each value comes from:

Column	Taken from (first one present)
`title`	`og:title`, `twitter:title`, `<title>`, JSON-LD `headline`, oEmbed `title`
`description`	`og:description`, `twitter:description`, `<meta name="description">`
`image`	`og:image:secure_url` / `og:image`, `twitter:image`, `<link rel="image_src">`, oEmbed `thumbnail_url`; for a URL that is itself an image, the URL
`imageWidth`, `imageHeight`, `imageAlt`	`og:image:width`, `og:image:height`, `og:image:alt` (or `twitter:image:alt`) as the page states them; the image is not downloaded
`siteName`	`og:site_name`, `<meta name="application-name">`, oEmbed `provider_name`
`favicon`, `appleTouchIcon`	`<link rel="icon">` (the size closest to 32 px), `<link rel="apple-touch-icon">`
`canonicalUrl`, `ogUrl`	`<link rel="canonical">`, `og:url`
`language`, `languageSource`	`<html lang>`, the `Content-Language` header, `og:locale`
`type`	`og:type`
`author`, `publishedTime`, `modifiedTime`	`<meta name="author">`, `article:author`, `article:published_time`, `article:modified_time`, `og:updated_time`, JSON-LD, oEmbed `author_name`
`openGraph`, `twitter`	every `og:` and `twitter:` tag, as written
`sources`	which of `openGraph`, `twitterCard`, `html`, `jsonLd` supplied the title, description, image or site name, plus `oembed` when the oEmbed endpoint was read, or `url` when the URL is itself an image

Relative URLs are resolved against the final URL (and <base href>), so image, favicon and canonicalUrl are always absolute.

Rows that are not charged

`status`	Meaning
`robots-disallowed`	The site's robots.txt does not allow automated reading of this address; it was not requested.
`robots-unreachable`	The site's robots.txt answered 5xx or could not be reached; following RFC 9309 the page was not requested.
`login-required`	HTTP 401, or the page redirects to a sign-in page.
`blocked`	The site refused this reader (HTTP 403, 429 after retries, 451) or showed a CAPTCHA or browser-check page (with any status; a check page is never asked again). A person with a browser may still see the page.
`not-found`	HTTP 404 or 410.
`http-error`	Another HTTP answer that is not a page (for example 400 or 405), or a broken redirect.
`unreachable`	The domain does not resolve, or the site did not answer.
`unreadable`	5xx after retries, an answer that kept stopping before the end of the page's `<head>`, or more than 8 redirects.
`no-metadata`	The page was read but has no title, description or image.
`not-html`	The address is a PDF, JSON, video or other file that is not a web page or an image.
`bad-input`	Not an http(s) URL, a URL with a user name or password, or an address in a private network.
`no-change`	Monitor mode: nothing changed (one summary row), or a change could not be confirmed by a second read.
`budget-reached`	The run reached the maximum total charge you set; the rest was not requested.

These rows keep the same columns as a preview row, with note explaining the reason and httpStatus, finalUrl and redirects filled in when known.

Pricing

Pay per event:

Preview returned: charged for each row with status: "ok". A URL that is itself an image (Content-Type image/*) is an ok row with image set to that URL and the other preview values (title, description, site name and so on) null.
Run start: charged once per run that returns at least one preview. In monitor mode it is charged once per run that read and compared at least one URL, whether or not anything changed. It is not charged when no URL could be previewed or the input could not be used.

A run whose maximum total charge has no room for the run start fee plus one preview requests nothing and is charged nothing. The current prices are on the Pricing tab.

Calling it from code

Synchronous call that returns the rows directly (replace <YOUR_TOKEN>):

curl -X POST "https://api.apify.com/v2/acts/neverempty~link-preview-api/run-sync-get-dataset-items?token=<YOUR_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://github.com/apify/crawlee"]}'

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('neverempty/link-preview-api').call({ urls: ['https://github.com/apify/crawlee'] });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0].title, items[0].image);

Monitor mode

Run the same list on a schedule with onlyChanges: true and a watchName. The first run returns every URL (changeType: "first-check"). Later runs return only URLs whose title, description, image, site name, canonical URL or type changed (changeType: "changed", with changedFields, previousTitle, previousDescription, previousImage and previousCheckedAt), and URLs added to the list ("new").

A change is reported only when a second read a moment later shows the same new value. A value that differs between the two reads (for example an image that rotates on every visit) is not reported and is not compared for that URL again.
Image URLs are compared without their query string, so signed or cache-busting CDN parameters do not count as a change.
Use a different watchName for each list you track on its own schedule, and avoid two overlapping schedules with the same watchName (the remembered state is saved after each run and Apify has no atomic update).

Measured on 2026-09-24

40 commonly shared URLs (news sites, GitHub, YouTube, Wikipedia, Spotify, Vimeo, PyPI, social networks, Japanese, French and German sites) in one run from Apify: 23 previews, 6 not requested because robots.txt disallows them (the big social networks), 7 refused or check pages, 2 not found, 2 sites that did not answer. The run took 54 seconds with 4 in parallel; peak memory 176 MB of 256 MB.

Limits

JavaScript is not run. Sites that write their tags only in the browser (single-page apps without server rendering) may return fewer values; the row says what the served HTML contains.
Pages behind a sign-in, and sites whose robots.txt disallows automated reading, are not read. Most large social networks disallow it.
Values are what the site serves to a reader in a US data center; a site may serve different content by country or language.
This is an independent tool and is not affiliated with any website it reads.

Support

Questions, a URL that returns something unexpected, or a missing tag: open an issue in the Issues tab with the URL and the run ID.

YouTube Most Replayed Scraper - Heatmap Peaks & Timestamps

neverempty/youtube-most-replayed-scraper

Find the most replayed moments of any YouTube video: the top peaks of YouTube's own Most replayed heatmap as timestamped links, the chapter each peak falls in and the caption line spoken there, plus the raw heatmap segments. Videos without a heatmap are listed free.

NeverEmpty

Open Graph Social Preview Auditor

glowing_glove/open-graph-social-preview-auditor

Audit URLs for Open Graph, Twitter Card, canonical, image, and share-preview gaps that hurt social and SEO campaigns.

Ushba Khan

Social Preview Card Verifier

nexgenwatch/social-preview-card-verifier

Verifies how a page will unfurl on social platforms. Reads Open Graph + Twitter Card tags, validates each declared image for reachability, MIME and pixel dimensions, and returns a per-platform readiness verdict with every fault named.

NexGen Watch

Social Preview Tags — Open Graph and Twitter cards per domain

artifact-machine/social-preview-enrichment

Checks whether each domain has the Open Graph and Twitter Card tags that control how its links look when shared. Finds sites whose links render blank on LinkedIn, Slack and X.

The Artifact Machine

Link Preview Metadata: Open Graph, Twitter Card, Favicon

jtpalms/link-preview-metadata

Get what a link unfurler shows for any URL: title, description, image, site name, every Open Graph and Twitter card tag, best favicon, theme color, language, author, publish date, JSON-LD types, oEmbed URL, H1 and word count. For app link previews and AI agents. USD 1 per 1,000 URLs.

JT Palms

RSS Feed Scraper - Atom & JSON Feed Reader, New Item Monitor

neverempty/rss-feed-scraper

Give it feed URLs or plain website URLs. It finds the feed a site advertises, reads RSS 2.0, RSS 1.0/RDF, Atom and JSON Feed in any encoding, and returns one row per item with title, link, dates, author, summary, full text, categories, image and enclosures. Monitoring mode returns only new items.

NeverEmpty

Link Preview API — OpenGraph & Metadata Unfurl Tool

parsebird/link-preview-api

High-volume OpenGraph, Twitter Card, and JSON-LD unfurl API. Extract title, description, image, canonical URL, favicon, feeds, hreflang, and schema.org data from any list of URLs.

ParseBird

Social Preview Validator - Open Graph & Twitter Cards

quanmatrix/social-preview-validator

Validate Open Graph, Twitter/X card, canonical, and social preview image health for public URLs in bulk, returning structured metadata, issue flags, and a 0-100 QA score.

Rafael Barreto Haddad

App Store App Details Tracker - Ratings, Star Counts & Versions

neverempty/app-store-app-details-tracker

For ASO teams, app marketers and competitor trackers: exact rating count, count per star (5 to 1), rating, version, update date, price, size and chart rank of App Store apps. Matched the live store page on 5 of 5 apps checked 2026-09-21. Monitoring mode returns only apps that changed.

NeverEmpty

Link Preview & Metadata Extractor

weeknds/link-preview-metadata-extractor

Extract Open Graph, Twitter Cards, meta descriptions, favicons, and all metadata from any URL. Structured link previews with canonical URLs, language detection, HTTP status codes. Pure HTTP, no browser needed. Perfect for SEO audits and content aggregators.