How does speakable markup select text to read aloud?

Definition: SpeakableSpecification identifies sections of an article or webpage for text-to-speech. Google documents two alternative content locators: cssSelector, which addresses page content with CSS selectors, and xPath, which addresses content using XPath values.

Key facts

• Google's Speakable documentation describes speakable as identifying sections within an article or webpage for text-to-speech playback.
• The speakable property can be repeated, and Google documents two content-locator choices: CSS selectors and xPaths.
• cssSelector addresses content in the annotated page, such as content identified by a class attribute. Google says to use either cssSelector or xPath, not both.
• xPath addresses content using xPaths while assuming an XML view of the content. It is the alternative to cssSelector within the same SpeakableSpecification.
• Schema.org defines SpeakableSpecification as indicating document sections highlighted as particularly speakable, typically through xpath or cssSelector.

How does cssSelector choose the content?

The cssSelector property contains CSS selector values that address content in the annotated page. Google's example uses an array with two class selectors, .headline and .summary, inside a SpeakableSpecification object:

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

Google's property table describes cssSelector as addressing content such as a class attribute. The values therefore point to page elements whose content is marked for the speakable property. Schema.org's cssSelector definition calls the value a CSS selector of a SpeakableSpecification or a WebPageElement.

How does xPath choose the content?

The xPath property is the other locator form. Google describes it as addressing content with xPaths while assuming an XML view of the content. Its example points to the document title and the content attribute of a description meta element:

"speakable": {
  "@type": "SpeakableSpecification",
  "xPath": [
    "/html/head/title",
    "/html/head/meta[@name='description']/@content"
  ]
}

The Google documentation requires a choice between cssSelector and xPath in this object. Schema.org's xpath definition describes an XPath of a SpeakableSpecification or WebPageElement. For a WebPageElement, it says that multiple matches within a page can form one conceptual web page element.

What does the selector list represent?

As of 8 September 2026, Google labels the Speakable feature beta and says that requirements or guidelines may change. The speakable property identifies sections suited to audio playback, so a locator list identifies the page sections that the specification marks for that purpose.

Google's technical guidance says to focus on key points instead of marking an entire article. Its content guidance calls for concise headlines or summaries, and recommends around 20 to 30 seconds of content per section, or roughly two to three sentences. The selector syntax chooses the elements; the content guidance describes what those elements should contain.

How can you check the markup?

Check the JSON-LD object for "@type": "SpeakableSpecification" and exactly one of the two documented locator properties. For CSS selection, compare each value with the page's class or other CSS-selectable structure. For XPath selection, compare each path with the XML-style document path it is intended to address. The two examples above show the property names, array form, and value formats documented by Google.

Related terms

• Speakable structured data
• Article structured data
• WebPage structured data
• Structured data
• JSON-LD

FAQ

Can cssSelector and xPath appear together?

No. Google's Speakable documentation says to use either property and not both in the same specification.

Can a page have more than one speakable specification?

Yes. Google says the speakable property can be repeated an arbitrary number of times. Each SpeakableSpecification can use one of the two locator forms.

Does the selector identify the whole article?

The selector identifies the addressed page content. Google advises focusing on key points rather than highlighting an entire article, and Schema.org describes SpeakableSpecification as indicating highlighted sections of a document.

Sources

1. Google Search Central, "Speakable (Article, WebPage) structured data (BETA)", 2026-09-08
2. Schema.org, "SpeakableSpecification", 2026-09-16
3. Schema.org, "cssSelector", 2026-09-16
4. Schema.org, "xpath", 2026-09-16

---

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