window.SpeechSynthesis is a singleton. You might, for instance, create a
SpeechSynthesisUtterance then call
window.SpeechSynthesis.speak(utterance) to speak it. You don't need a
SpeechSynthesis instance to do this.
But if you look at SpeechSynthesis in web_sys - Rust, it appears to need
How am I meant to call this API?
SpeechSynthesisUtterance is easy as it has a
new method, but I don't see something similar for