Page metadata - Context.dev

Every POST /web/scrape response includes a metadata object, whatever you enable in formats. It describes the page itself rather than one output: the title, description, language, social preview tags, headings, and JSON-LD blocks found in the page HTML. Fields are omitted when not found, so read each one defensively.

Read page metadata

Send this body to POST /web/scrape. See the Quickstart for authentication and SDK setup.

{
  "url": "https://example.com",
  "formats": {
    "markdown": true
  }
}

Abridged response:

{
  "metadata": {
    "title": "Example Domain",
    "description": "This domain is for use in illustrative examples in documents.",
    "language": "en",
    "canonicalUrl": "https://example.com/",
    "siteName": "Example",
    "image": "https://example.com/og-image.png",
    "favicon": "https://example.com/favicon.ico",
    "openGraph": {"title": "Example Domain", "type": "website"},
    "twitter": {"card": "summary"},
    "alternates": [{"href": "https://example.com/fr/", "hreflang": "fr"}],
    "headings": [{"level": 1, "text": "Example Domain"}],
    "jsonLd": [{"@context": "https://schema.org", "@type": "WebSite", "name": "Example"}],
    "additionalMeta": {"viewport": "width=device-width, initial-scale=1"}
  }
}

The Scrape overview reads metadata.title next to markdown.data; the same object is there for every format, including screenshots and bytes.

Fields

Field	Notes
`title`	String. Best title extracted from the page.
`description`	String. Best description extracted from standard, Open Graph, or Twitter metadata.
`language`	String. Language extracted from `html lang` or language meta tags.
`keywords`	Array of strings. Keywords extracted from the page’s keywords meta tag.
`canonicalUrl`	String. Resolved canonical URL, when present.
`author`	String. Author metadata, when present.
`siteName`	String. Site or application name from page metadata.
`image`	String. Primary resolved preview image from Open Graph, Twitter, or image metadata.
`favicon`	String. Resolved favicon URL, when present.
`publishedTime`	String. Published timestamp or date from page metadata, when present.
`modifiedTime`	String. Modified timestamp or date from page metadata, when present.
`robots`	String. Robots meta directive, when present.
`openGraph`	Object. Open Graph metadata with the `og:` prefix removed and keys camel-cased. Each value is a string or an array of strings.
`twitter`	Object. Twitter card metadata with the `twitter:` prefix removed and keys camel-cased. Each value is a string or an array of strings.
`alternates`	Array of objects. Resolved alternate links from `link rel=alternate` tags; see Alternates.
`headings`	Array of objects. Up to 500 h1–h6 headings in document order, before content filtering; see Headings.
`jsonLd`	Array of objects. JSON-LD structured data blocks parsed from the page.
`additionalMeta`	Object. Additional non-social meta tags not promoted to top-level metadata fields. Each value is a string or an array of strings.

Alternates

Field	Notes
`href`	String, always present. Resolved alternate URL.
`hreflang`	String. Language or locale for the alternate URL, when present.
`type`	String. Alternate resource MIME type, when present.
`title`	String. Alternate resource title, when present.

Headings

Field	Notes
`level`	Integer from 1 to 6, from `h1` to `h6`.
`text`	String. Heading text with whitespace collapsed, truncated to 1,000 characters.

Headings are collected before content filtering runs, so mainContentOnly and selector filters do not change them.

Scrape overview: the full response shape and the outputs each format adds.
Crawl page metadata: the same fields on every crawled page, plus crawl-specific ones.
Content filtering: what the filters change and what they leave alone.
Freshness and caching: when a response is served from cache.