Speakable structured data

The Citead team, updated

In short. Speakable marks passages for text-to-speech; Google documents its use for English news answers on Google Home devices in the U.S.

Definition: Speakable structured data identifies parts of an article or web page as suitable for text-to-speech, the conversion of written text into spoken audio. Schema.org defines the speakable property for highlighting passages that are especially useful when read aloud.

Key facts

  • As of October 2026, Google's speakable documentation labels the feature beta and limits availability to English content for U.S. users with Google Home devices set to English.
  • Schema.org's property definition accepts a URL or a SpeakableSpecification as a value on Article and WebPage.
  • Google's AI features guidance says AI Overviews and AI Mode require no special Schema.org structured data.

What use does Google document?

Google's implementation guide describes topical news answers on smart speakers. Google Assistant reads marked passages, attributes the source and sends the full article URL to the user's mobile device through the Assistant app.

How does the markup identify a passage?

A content locator identifies the part of a document to read. Schema.org's speakable definition describes references to element IDs, CSS selectors and XPath expressions. An ID reference points to an identified section through a URL fragment. A CSS selector targets elements through patterns such as class names; XPath addresses locations in a document viewed as XML.

The supporting SpeakableSpecification type holds selectors or paths. It describes the selected sections; the speakable property associates that specification with the article or page.

Google's CSS selector example supplies this property fragment for a containing Article or WebPage object:

"speakable": {
  "@type": "SpeakableSpecification",
  "cssSelector": [".headline", ".summary"]
}

The selectors address elements with the headline and summary classes. Google requires either cssSelector or xPath for its implementation, with only one locator method used.

Which text is suitable for speech?

Google's content guidelines call for concise headlines or summaries and recommend around 20-30 seconds per section, roughly two to three sentences. Key points are preferred over an entire article. Datelines, photo captions and source attributions should be excluded from marked passages because they can confuse listeners.

FAQ

Is speakable required for Google AI answers?

Google's AI features documentation says you need no special markup for AI Overviews or AI Mode. Eligibility as a supporting link requires an indexed page that can appear in Google Search with a snippet. The same documentation advises keeping structured data consistent with visible page text.

Can a page identify more than one passage?

Yes. Schema.org's speakable definition allows the property to repeat and includes an example with separate URL fragments for a heading and a summary.

Sources

  1. Google Search Central, "Speakable (Article, WebPage) structured data (BETA)", updated 8 September 2026
  2. Schema.org, "speakable", accessed 17 October 2026
  3. Schema.org, "SpeakableSpecification", accessed 17 October 2026
  4. Google Search Central, "AI features and your website", updated 10 December 2025

Check how a page reads to answer engines. Open the GEO score analyzer

Test a page and see what to fix first

The analysis usually takes less than 10 seconds and needs no account.

400,000+ pages scored in the public benchmark