speechstart

Speechstart is an event name used in several speech recognition interfaces to indicate the onset of human speech during a recognition session. It is not part of a single standardized specification, and implementations vary across platforms and libraries. In environments that expose it, the event fires when the system first detects speech from the user, typically after the recognition engine begins listening.

Applications commonly use speechstart to trigger user interface cues, such as highlighting the active microphone, starting a timer, or initiating preprocessing steps for incoming audio. It generally occurs before any transcription results are produced and may be followed by events like speechend or result. In code, it can often be listened for as an event named speechstart or onspeechstart, using standard event listening patterns.

Usage notes emphasize that support for speechstart is not universal. Developers should check for API availability before relying on it, since some browsers implement related events with different names or do not expose this event at all. Privacy and security considerations apply, as detecting speech activity can reveal when a user is actively speaking or present, even if transcription is not performed.

Related concepts include speechend, which signals the end of spoken input, and result or nomatch events, which convey transcription outcomes. Speechstart is commonly discussed in the context of the Web Speech API and other speech recognition libraries, where it forms part of the event-driven lifecycle that governs listening, processing, and interpreting spoken language.