Paste any URL, get a SpeakableSpecification JSON-LD block that marks the headings and direct-answer paragraphs AI assistants should read aloud. Selects answer-shaped content automatically, lets you override.
We fetch the HTML, find H2/H3 headings and the first paragraph after each, then build a Speakable block targeting those elements via CSS selectors. You can remove or keep individual matches before copying.