Built by Rankite, the SEO team behind Swordfish AI's +400% revenue and Zluri's +45% organic growth. See the case studies
Speakable structured data tells voice assistants and AI systems which parts of your page are written to be read aloud on their own, without the surrounding navigation, sidebars or unrelated content. It was introduced by Google for Assistant-powered news reading on smart speakers, and the underlying idea, flagging a short, self-contained, quotable passage, is increasingly useful for any system trying to extract a clean summary from a page.
The speakable property points to one or more sections of a page, most commonly by CSS selector, that are suitable for text-to-speech playback. It is added as a SpeakableSpecification object, either inside your page's existing WebPage or Article schema, or as its own JSON-LD block like the one this tool generates, with a cssSelector array listing each section you want included.
Google's guidance points to roughly 20 to 30 seconds of spoken audio per section, which in practice means about two to three well-written sentences. A headline plus a tight, factual summary paragraph is the classic combination. Sections that run much longer risk being truncated or skipped entirely by the systems that consume this markup.
Speakable supports both CSS selectors and XPath expressions as ways to point at content, but Google's documentation recommends choosing one method and sticking with it rather than mixing the two inside a single specification. CSS selectors are the more common, easier to test choice, and what this tool outputs by default.
Get a free, no-obligation SEO audit and a 30-minute strategy session. We'll show you exactly where the growth is hiding.
Fill out the form and we'll get back to you within one business day. Prefer email? Write to us directly at contact@rankite.com.