metadata

Unified metadata from Open Graph, Twitter Cards, JSON-LD, and the HTML itself, normalized into one predictable shape:

const { title, description, image } = await microlink.metadata('https://vercel.com')

It resolves to the data fields object — title, description, lang, author, publisher, image, logo, date, and url — with every image field expanded into an asset object carrying url, type, width, height, size, and size_pretty:

const { image, logo } = await microlink.metadata('https://github.com/microlinkhq')

console.log(image.url, image.width, image.height, image.size_pretty)
console.log(logo.url)

Fields that cannot be detected come back as null, so destructuring is always safe.

Options

It has no method-specific options: every key is a top-level API parameter, so the whole shared options set applies. A few of them are especially useful here.

meta <object> — restricts detection to specific fields, or excludes some, to make the call cheaper.
palette <boolean> — adds dominant colors and accessible color pairs to every image field (default: false).
video <boolean> — adds a video field with the primary video of the page as a direct asset (default: false).
audio <boolean> — adds an audio field with the primary audio track as a direct asset (default: false).
prerender <boolean> | <string> — forces a headless browser for client-side rendered pages whose metadata only exists after JavaScript runs (default: 'auto').

Only detect the fields you need:

const { title, author } = await microlink.metadata('https://x.com/microlinkhq', {
  meta: { title: true, author: true }
})

Add accessible color pairs to every image, ready to theme a preview card:

const { image } = await microlink.metadata('https://vercel.com', {
  palette: true
})

console.log(image.palette, image.background_color, image.color)

Get the primary video along with the rest of the metadata, instead of a separate video call:

const { title, video } = await microlink.metadata('https://www.youtube.com/watch?v=9P6rdqiybaw', {
  video: true
})

console.log(video.url, video.duration_pretty)

Extending the result

Combine metadata with your own extraction rules when the normalized fields are not enough. extract returns only your fields; to get both at once, pass the rules as data and the normalized fields ride along:

const { title, price } = await microlink.metadata('https://example.com/product', {
  data: {
    price: { selector: '.price', attr: 'text', type: 'number' }
  }
})

See the metadata guide for choosing fields, page preparation, caching, and private pages.