OptionalafterOptionalafterFire the interrupt only after the adapter's streaming transcript has emitted N words.
OptionalcontentOptional content to send as the interruption (string text or audio bytes/path).
OptionalwaitBounded wait for the agent to start speaking before firing the interrupt.
TIME-based trigger (PRD §4.4
interrupt(after=2.0, …)): wait this many seconds after the agent starts speaking before injecting the interruption. Equivalent toagent({wait:false}) + sleep(after) + user. Mutually exclusive withafterWords(which takes precedence if both set).