Speakable Structured Data: Complete Implementation Guide
Speakable structured data identifies sections of a web page that are best suited for audio playback using text-to-speech (TTS) engines, particularly Google Assistant. By marking up specific portions of your content as speakable, you help voice assistants select the most relevant and concise passages to read aloud in response to user queries. This is especially valuable for news publishers and content-heavy sites aiming to capture the growing voice search audience. Currently, Speakable is available in English for users in the United States, with broader language and region support expected over time.
Why Use Speakable Schema?
- Position your content for voice search results delivered by Google Assistant and other TTS-enabled platforms
- Control exactly which sections of your page are read aloud, ensuring the most relevant information is presented
- Gain a competitive advantage in voice search, an increasingly important channel as smart speaker adoption grows
- Improve accessibility for users who rely on screen readers or audio-based browsing
- Future-proof your SEO strategy as voice-first interfaces become more prevalent across devices
Required Properties
These properties are required by Google for Speakable structured data to be eligible for rich results.
| Property | Description |
|---|---|
| cssSelector or xpath | You must specify either cssSelector or xpath to identify the speakable sections of the page. cssSelector uses CSS selector syntax, while xpath uses XPath expressions. |
| cssSelector | A CSS selector that points to the HTML elements containing the content suitable for TTS. For example, '.article-summary' or '#headline'. |
| xpath | An XPath expression identifying the speakable content. An alternative to cssSelector — use whichever is more convenient for your markup structure. |
Recommended Properties
Adding these properties improves your chances of appearing in rich results and provides more context to search engines.
| Property | Description |
|---|---|
| name (on parent article) | The headline or title of the article containing speakable content. Helps Google understand the context of the speakable sections. |
| datePublished (on parent article) | The publication date of the content in ISO 8601 format. Voice results favour fresh, timely content. |
| image (on parent article) | An image associated with the article. While not read aloud, it provides context for Google's understanding of the content. |
| author (on parent article) | The author of the article. Attribution supports E-E-A-T signals that Google considers when selecting voice results. |
JSON-LD Example
Here is a complete Speakable structured data example you can use as a starting point. Or use our Speakable generator to create one automatically.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "WebPage",
"name": "Spring Budget 2026: Key Announcements Summary",
"speakable": {
"@type": "SpeakableSpecification",
"cssSelector": [
".article-headline",
".article-summary"
]
},
"url": "https://example.com/news/spring-budget-2026-summary",
"mainEntity": {
"@type": "NewsArticle",
"headline": "Spring Budget 2026: Key Announcements Summary",
"datePublished": "2026-03-10",
"author": {
"@type": "Person",
"name": "James Thornton"
},
"publisher": {
"@type": "Organization",
"name": "UK Finance Daily",
"logo": {
"@type": "ImageObject",
"url": "https://example.com/images/logo.png"
}
},
"image": "https://example.com/images/budget-2026.jpg"
}
}
</script>How to Implement Speakable Schema
- 1Identify the concise, self-contained sections of your page that would make sense when read aloud — typically headlines and summary paragraphs of 2-3 sentences.
- 2Add semantic CSS classes or IDs to the HTML elements containing your speakable content, such as 'article-headline' or 'article-summary'.
- 3Create a JSON-LD block using the WebPage type with a speakable property containing a SpeakableSpecification that references your CSS selectors.
- 4Ensure the speakable content is concise and informative — Google recommends roughly 20-30 seconds of audio per speakable section, which equates to approximately 2-3 sentences.
- 5Validate the markup using Google's Rich Results Test and check for any warnings related to the speakable specification.
- 6Submit your page for re-indexing via Google Search Console and monitor for any speakable-related enhancements in the Search Appearance report.
Common Mistakes to Avoid
- Marking up the entire article body as speakable — Google expects concise, summary-level content that works well as a standalone audio snippet, not full articles.
- Using speakable on non-news or non-informational content where voice search results are unlikely to surface, reducing the value of the implementation.
- Specifying CSS selectors that do not match any elements on the page, resulting in empty speakable content that Google cannot process.
- Forgetting that Speakable is currently limited to English-language content for US users — implementing it for other languages or regions will not yet produce results.
- Neglecting to include a parent NewsArticle or Article type alongside the speakable specification, which reduces Google's ability to contextualise the content.
Testing & Validation
Before deploying your Speakable structured data, validate it using these tools:
- 1Google Rich Results Test — Paste your page URL or code snippet to check if your structured data is eligible for rich results.
- 2Schema Markup Validator — Validates your JSON-LD syntax against the schema.org specification.
- 3Google Search Console — After deployment, monitor the Enhancements report for any errors or warnings.
Frequently Asked Questions
Is Speakable structured data available outside the United States?
As of now, Speakable is only supported for English-language content targeting users in the United States. Google has indicated plans to expand support to additional languages and regions, but no specific timeline has been confirmed. Implementing it now prepares your site for when support broadens.
What type of content works best with Speakable markup?
News articles, weather updates, sports results, and factual summaries perform best. The content should be concise (2-3 sentences per speakable section), factual, and self-contained so it makes sense when read aloud without visual context. Avoid content that relies on images, charts, or links to convey meaning.
Should I use cssSelector or xpath for my speakable content?
CSS selectors are generally recommended as they are more widely understood and easier to maintain. Use xpath only if your content structure requires more precise targeting that CSS selectors cannot achieve. You must use one or the other — not both — for each SpeakableSpecification.
Does Speakable structured data affect traditional search rankings?
Speakable markup does not directly influence traditional web search rankings. Its purpose is to make your content eligible for voice search results via Google Assistant. However, being selected as a voice result increases brand exposure and can drive traffic through follow-up queries.
Can I add Speakable markup to non-article pages?
Technically, Speakable can be applied to any WebPage type. However, Google currently prioritises news and informational content for voice results. Applying it to product pages or landing pages is unlikely to produce results at this time, though this may change as the feature evolves.
Ready to Generate Your Speakable Schema?
Use our free generator to create valid Speakable structured data in seconds.
Open Speakable Generator