Speakable schema
Speakable schema is a Schema.org structured-data property (using the SpeakableSpecification type) that marks the specific sections of a web page best suited to be read aloud by text-to-speech engines, such as voice assistants and answer engines. You point it at the most quotable parts of your page using an HTML element id, a CSS selector, or an XPath.
What Speakable schema actually does
Speakable is a property on Schema.org's structured data, applied via the SpeakableSpecification type. Instead of describing your whole page, it points machines at the exact passages worth speaking aloud — typically a concise headline and a two-to-three-sentence summary. You attach it three ways:
- CSS selector — target an element by its class or id, e.g. the summary paragraph
- XPath — address a node by its path in the document
- id reference — point at an element's HTML id directly
Google recommends roughly 20 to 30 seconds of content per speakable section (about two to three sentences), and warns against marking up datelines, photo captions, or source attributions that sound confusing without context. It is a sibling signal to general schema markup, aimed specifically at audio and voice surfaces.
Why it matters for GEO and voice
Speakable was introduced so voice assistants and answer engines could lift a clean, self-contained snippet from your page rather than guessing. That same instinct — give the machine the most quotable, context-free passage — is the core of generative engine optimization. When ChatGPT, Perplexity, or AI Overviews assemble an answer, pages that surface a tight, standalone summary are easier to pull and easier to attribute.
Treat Speakable less as a guaranteed rich result and more as discipline: it forces you to write a crisp, liftable summary near the top of every page. That structure improves your odds of an AI citation and a featured snippet even on engines that never parse the speakable property itself. Validate any markup in a structured-data tester — Google's news Speakable feature has shifted over time, so confirm current behavior before claiming a specific result.
How an approval-gated growth team handles it
Speakable markup is one small lever inside a broader citation strategy that also covers entity clarity, topical authority, and clean summaries. Ceres is a managed AI growth team where a GEO Strategist (one of 11 specialists) drafts that schema and the surrounding GEO changes, then routes them to you. You stay the boss: every outbound or published change is approval-gated, so a human reviews the markup before it ships. You can see where your pages stand today with the free GEO audit.
FAQ
- What is Speakable schema?
- Speakable schema is a Schema.org property (the SpeakableSpecification type) that flags the sections of a page best suited to be read aloud by text-to-speech, identified by a CSS selector, XPath, or element id. It tells voice assistants and answer engines which passage to lift.
- Does Speakable schema help with AI search and GEO?
- Indirectly, yes. The property itself targets text-to-speech, but the practice of marking a tight, self-contained summary forces the kind of quotable, context-free writing that AI engines like ChatGPT and Perplexity prefer to cite. It complements broader GEO work like entity clarity and topical authority.
- How do I add Speakable schema to a page?
- Add a SpeakableSpecification block to your page's JSON-LD or microdata and point it at your summary using cssSelector, xpath, or an id reference. Keep each section to about 20 to 30 seconds of audio (roughly two to three sentences) and validate it in a structured-data testing tool before publishing.
An AI growth team that runs this for you
Ceres is a managed AI marketing team — you approve what ships. 14-day free trial, from $19/month.