Meta tags and sitemap hints

generateMeta

With generateMeta: true, the integration derives <meta> and <link> elements from the primary schema of the page and inserts them into <head>. This runs as middleware after the page has fully rendered, so it sees every schema on the page regardless of where the components are placed, including components that await data first.

• Primary schema: the most page-specific schema wins (Article types, Product, Recipe, VideoObject, Event, JobPosting, SoftwareApplication, ProfilePage, FAQPage, WebPage) before site-wide LocalBusiness and WebSite schemas. It drives title, description, image and og:type. An @type array such as ['BlogPosting', 'LearningResource'] matches by each of its types.
• Page schemas: page-level tags (robots, hreflang alternates, the canonical URL) come from the schemas describing the page itself, WebPageSchema and its variants, whichever content schema is primary. With several of them, such as a FAQPage and a WebPage, settings from each count and robots directives from all of them are kept. A noindex WebPage next to an Article stays noindex.
• Your tags win: tags the page already defines (same name or property, canonical, hreflang) are kept and not duplicated. If the page sets its own og:image, the schema image’s width, height, type and alt are left out, and so is twitter:image (Twitter falls back to og:image).
• Canonical and og:url are built from siteUrl and the page path, or taken from a WebPageSchema url or an article’s mainEntityOfPage. The url of an event, business or product is never used as the page’s canonical. If the page sets its own canonical link, og:url uses that URL too, so both always agree.
• Rewrites: after a middleware rewrite (next('/target') or context.rewrite()), the tags come from the schemas of the rewrite target, and the page path for the canonical is the target’s path, the same path the components build their page URLs from. To point an alias elsewhere, pass url to WebPageSchema.
• A page with a primary schema but no </head> fails the build.

HTML responses are buffered by the middleware rather than streamed.

Generated tags

• Canonical, description, robots, author, reading-time (non-standard)
• Alternates: <link rel="alternate" hreflang="…"> from alternates or from workTranslation (url and inLanguage of each translation)
• Open Graph: og:title, og:description, og:image with width, height, type and alt, og:url, og:type, og:site_name, og:locale, and for articles article:published_time, article:modified_time, article:author (author profile URLs; the name goes into <meta name="author">), article:section, article:tag; for videos og:video (plus og:video:secure_url with an embedUrl) and video:duration in seconds
• Twitter: twitter:card, twitter:title, twitter:description, twitter:image, twitter:image:alt, twitter:site, twitter:creator

Relative image URLs in the meta tags are resolved to absolute URLs with siteUrl the way a browser would: /image.jpg against the origin, image.jpg against the site path. siteName, locale, twitterSite and twitterCreator from the integration options fill the site-wide tags.

Example

The ArticleSchema of the Quickstart on /blog/hello-astro/, with structuredData({ generateMeta: true, siteName: 'Example Blog' }) and site: 'https://example.com', adds these tags to <head>:

<link rel="canonical" href="https://example.com/blog/hello-astro/">
<meta name="description" content="Our first post.">
<meta name="author" content="Ada Lovelace">
<meta property="og:title" content="Hello, Astro">
<meta property="og:description" content="Our first post.">
<meta property="og:url" content="https://example.com/blog/hello-astro/">
<meta property="og:image" content="https://example.com/img/hello.jpg">
<meta property="og:site_name" content="Example Blog">
<meta property="og:type" content="article">
<meta property="article:published_time" content="2026-10-01">
<meta property="article:modified_time" content="2026-10-01">
<meta name="twitter:title" content="Hello, Astro">
<meta name="twitter:description" content="Our first post.">
<meta name="twitter:image" content="https://example.com/img/hello.jpg">
<meta name="twitter:card" content="summary_large_image">

Robots and alternates

WebPageSchema accepts robots (e.g. 'noindex') and alternates ({ href, hreflang }[]). Both only drive meta tags and are not part of the JSON-LD. They apply even when another schema on the page, such as an article, is the primary one:

<WebPageSchema title="Draft" robots="noindex" />
<ArticleSchema title="Draft post" … />

The page gets the article’s title and Open Graph tags plus <meta name="robots" content="noindex">.

Sitemap hints

Every component, and the generic <Schema>, accepts changefreq and priority:

<ArticleSchema … changefreq="weekly" priority={0.8} />

They are written as data attributes on the JSON-LD <script> tag (on the @graph script in graph mode), not into the JSON-LD. A page has one sitemap entry, so all schemas on a page must agree; conflicting values, a changefreq outside the sitemap protocol values or a priority outside 0–1 fail the build. Post-build sitemap generators such as @casoon/astro-site-files can read these attributes from the HTML. Without such a generator they have no effect.