BEACON Guide

Speakable schema markup: the passages AI reads aloud.

Speakable schema (the SpeakableSpecification type) tells voice assistants and AI engines which parts of a page are the clean, spoken-answer summary. As AI Overviews and assistants increasingly answer out loud, marking your best answer passages makes you the source they quote. Here’s what it is and exactly how to add it.

Start free Run a free check
Built for spoken answersSpeakable identifies the sentence(s) an assistant should read back — your TL;DR, definition or direct answer.
It’s a JSON-LD propertyA `speakable` block of type SpeakableSpecification inside your Article or WebPage schema, pointing at content by CSS selector or XPath.
cssSelector or xpathReference the element holding your answer (e.g. ".answer-summary") so engines know precisely what to speak.
Pairs with FAQ & ArticleSpeakable works best layered onto Article/BlogPosting and FAQPage schema — the formats AI already favours.
Keep it shortSpeakable passages should be a couple of concise sentences — voice answers are read in full, so brevity wins.
Beacon detects itBeacon’s audit now flags whether your pages expose Speakable schema and recommends where to add it.

What is Speakable schema (SpeakableSpecification)?

Speakable is a schema.org property that marks sections of a page suitable for text-to-speech. You add it as a `speakable` property — of type SpeakableSpecification — on an Article, NewsArticle or WebPage, and point it at the part of the page that contains the spoken-friendly answer. Google introduced it for news voice results, and the same signal now helps AI assistants and answer engines identify the passage worth quoting.

The JSON-LD to add

Add a speakable block to your existing Article/WebPage schema. Using a CSS selector is the simplest approach:

  • "@type": "WebPage" (or Article/NewsArticle)
  • "speakable": { "@type": "SpeakableSpecification", "cssSelector": [".answer-summary", "h1"] }
  • Point the selector at your direct-answer paragraph and the headline
  • Validate with Google’s Rich Results Test and the Schema.org validator

Google speakable markup vs. general schema

Regular schema (Article, FAQ, Product) tells engines what a page is; speakable tells them which words to say. They are complementary — speakable has no effect without a host Article/WebPage schema, so add the structural schema first, then layer speakable onto your answer passages.

Does speakable schema help AI search?

It’s a low-risk, high-signal GEO move. By explicitly marking the concise answer, you make your content trivially extractable for AI Overviews, ChatGPT, Perplexity and voice assistants — exactly the passage they want to read back. Combine it with a TL;DR answer block at the top of the page and FAQ schema for the strongest effect. Beacon’s free scan reports whether your pages already expose speakable schema.

FAQ

What is speakable schema markup?

A schema.org property (SpeakableSpecification) added to Article or WebPage JSON-LD that marks which passages are suitable to be read aloud by voice assistants and AI answer engines.

What is Google speakable markup?

It’s Google’s name for the same SpeakableSpecification schema — originally used for news voice results on Google Assistant, and a useful signal for AI answers identifying your best spoken summary.

How do I add speakable schema?

Add a `speakable` property of type SpeakableSpecification to your page’s Article/WebPage JSON-LD, using cssSelector or xpath to point at your answer summary and headline, then validate with Google’s Rich Results Test.

Does speakable schema markup help with AI search?

Yes — it makes the exact passage you want quoted machine-readable, improving the odds that AI Overviews, ChatGPT, Perplexity and voice assistants read and cite your content.

How do I check if my site has speakable schema?

Run Beacon’s free AI Visibility check — the audit detects Speakable (and Article/FAQ) schema on your pages and recommends where to add it.

Check your schema for AI & voice — free

See whether your pages expose Speakable and Article schema in seconds.

Start free Run a free check