Listening continuously and separating user phases by silence?

[scpedicini]

I have turned on continuous to true and as far as I can tell it continues to update the transcript string with the running transcript, and works great. When I've used the webkitSpeechRecognition in the past, I've been able to subscribe to onresult and then look for a var called isFinal.

This would allow me to listen continuously but "divide" each user's spoken stream into separate segments. For example, if a user says:
"turn off the lights please".

If they are silent for a while, the results will show that isFinal is set to true along with a confidence value. I can then act accordingly on that transcript, but all the while we are still continuously listening, and they may then say:
"report the weather"

Again followed by silence, etc.

How can I achieve this with react-speech-recognition?

[scpedicini]

@JamesBrill

FYI: I think I've got a workaround in place, just monitor whenever the interimTranscript field gets cleared out and record the last valid data from it.

const transcriptRef = useRef(interimTranscript);

useEffect(() => {
    if(interimTranscript === '' && transcriptRef.current.trim().length > 0) {
        const user_speech = transcriptRef.current.trim();
        resetTranscript();

        console.log(`useEffect - user said: ${user_speech}`);
    }
    transcriptRef.current = interimTranscript;

}, [interimTranscript, resetTranscript]);