Glossary

Speakable

A Schema.org property used to identify sections of a webpage or article that are most suitable for text-to-speech conversion by voice assistants, allowing publishers to control the audio presentation of their content.

Get in touch Learn more

ML engineer managing model versions on laptop, version history visible, technical Git-like workflow.

SCHEMA.ORG PROPERTY

What is Speakable?

Speakable is a Schema.org property that identifies sections of a webpage or article most suitable for text-to-speech conversion by voice assistants, giving publishers direct control over the audio presentation of their content.

Speakable is a structured data property within the Article and WebPage Schema.org types that allows publishers to explicitly mark up specific text segments—identified by CSS selectors or XPath—as the optimal content for text-to-speech (TTS) conversion. This specification directly addresses the gap between visual web content and audio consumption, ensuring that voice assistants like Google Assistant read only the core narrative rather than extraneous navigation elements, advertisements, or footer text.

Implementation requires wrapping target content in a SpeakableSpecification type, using either xpath or cssSelector properties to pinpoint the exact DOM nodes. Search engines, particularly Google, use this signal to power audio news briefings and smart speaker responses, prioritizing content that is clearly marked up. Proper deployment enhances content accessibility and ensures a publisher's editorial intent is preserved when their articles are consumed in screenless, voice-first environments.

Voice Assistant Optimization

Key Characteristics of Speakable

The Speakable property allows publishers to explicitly define which sections of an article are most suitable for text-to-speech conversion, ensuring a high-quality audio experience for users interacting via voice assistants.

Core Definition and Mechanism

Speakable is a Schema.org property applied to sections of a webpage, typically an Article or WebPage, to identify content optimized for text-to-speech (TTS) conversion. It uses XPath or CSS selectors to pinpoint specific HTML elements, allowing voice assistants like Google Assistant to read aloud only the most relevant parts, bypassing navigation, ads, and footers.

Technical Implementation via JSON-LD

Implementation requires a SpeakableSpecification type within a WebPage or Article schema. The xpath or cssSelector property points to the DOM nodes containing the speakable text.

xpath: /html/head/title for the headline
cssSelector: .article-body for the main content
Critical Rule: The SpeakableSpecification must not reference content that is invisible to the user or hidden via CSS.

Content Selection Best Practices

Selecting the right content is critical for user experience. The speakable text should be a concise, standalone summary or the core narrative.

Prioritize: The article headline and the first 2-3 paragraphs.
Exclude: Author bios, related links, sidebar widgets, and image captions.
Goal: Provide a complete, logical audio snippet that makes sense without visual context, typically lasting 20-30 seconds.

Relationship to Google Assistant

Google explicitly uses the Speakable markup to power its 'Read It' feature on Google Assistant. When a user asks for articles on a topic, the Assistant can read the speakable section aloud and attribute the source. Without this markup, the Assistant must algorithmically guess the main content, often resulting in a poor audio experience that includes irrelevant text.

Distinction from Other Schema Types

Speakable is often confused with other voice-related concepts but serves a distinct purpose.

vs. ReadAction: ReadAction indicates an app can read a document; Speakable specifies which part of a webpage to read.
vs. AudioObject: AudioObject is for a pre-recorded audio file; Speakable is for dynamic TTS generation.
vs. SSML: SSML controls pronunciation; Speakable controls content selection.

Impact on Generative Engine Optimization

As search shifts toward answer engines and voice interfaces, Speakable markup becomes a critical authority signal. By defining a clean, factual audio snippet, publishers provide a definitive source for voice answers. This structured approach helps generative models cite the content accurately, reinforcing the publisher's position as a high-confidence source in a voice-first ecosystem.

Enabling Efficiency, Speed & Accuracy

Intelligent Analysis, Decision & Execution

We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.

Talk to Us

Search across company data

Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.

Useful when people spend too long searching or get different answers from different systems.

Enterprise searchRAGPermissions

Automate internal workflows

Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.

Useful when repetitive work moves across multiple tools and teams.

AI agentsWorkflow automationGovernance

Add AI to products and internal tools

Build assistants, guided actions, or decision support into the software your team or customers already use.

Useful when AI needs to be part of the product, not a separate tool.

AI integrationDecision supportModel routing

SPEAKABLE SCHEMA

Frequently Asked Questions

Clear, concise answers to the most common technical questions about implementing the Speakable property for voice assistant optimization.

Speakable is a Schema.org property used to identify specific sections of a webpage or article that are most suitable for text-to-speech (TTS) conversion by voice assistants and screen readers. It works by allowing publishers to explicitly mark up content—typically the headline and a concise summary—using speakable within a SpeakableSpecification type. When a voice assistant like Google Assistant processes the page, it prioritizes the marked-up content for audio playback rather than attempting to read the entire page, which often includes navigation, ads, and other non-narrative elements. This gives publishers direct control over the audio presentation of their content, ensuring the spoken version is coherent, concise, and contextually appropriate. The speakable property accepts SpeakableSpecification as its value, which in turn uses cssSelector or xpath to pinpoint the exact DOM elements containing the speakable text.

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.

Limited slotsGet a Free AI Consultation

How We Work

Custom AI workflows for your Business

One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.

Talk to Us

Speakable

What is Speakable?