CKTIS / CSS
speak — controlling whether text is spoken
The CSS speak property determines whether text participates in speech rendering. Distinguish the CSS Speech syntax auto | never | always from the historical CSS2 syntax, and verify actual spoken output in the target environment.
Version: CSS modules; structured CSSWG indexes reconciled with pinned MDN BCD; Chromium 152 scope gate, per-example verification still required
Syntax
speak: auto | never | always;speak controls whether text participates in speech rendering. It applies to all elements, is inherited, and initially has the value auto.
With auto, display: none suppresses speech; otherwise, visibility determines the result. never suppresses the element, including its pauses and audio cues. A descendant can override this value. always enables speech regardless of display and ancestors’ speak values.
The informative CSS2 appendix describes the older normal | none | spell-out, initially normal. Do not mix these grammars: in CSS Speech, pronunciation behavior, including spelling out letters, belongs to the separate speak-as property.
The example illustrates intended semantics, not verified browser or screen-reader support. Do not base essential accessibility decisions solely on speak; verify actual spoken output.
Examples
<!doctype html>
<html lang="en">
<meta charset="utf-8">
<title>speak example</title>
<style>
p { speak: auto; }
.omit { speak: never; }
.restore { speak: always; }
</style>
<p>Alpha.</p>
<p class="omit">Beta. <span class="restore">Gamma.</span></p>
</html>Result
Visual text: Alpha. Beta. Gamma. Intended speech in a conforming CSS Speech renderer: Alpha. Gamma.
The displayed code and declared result were verified in an isolated environment using the method recorded in the verification evidence for this exact version.