Ooops, my big bad, I'll add a memo regarding this!

I couldn't make it work in Firefox yet: FF has a limited support for SpeechRecognition, that is enabled only via a flag. Even then it didn't work with my example. I'll dig into it further!

Meanwhile, to lessen the disappointment:

You can try this example in desktop/android chrome

OR

In FF open about:config and turn on these flags media.webspeech.recognition.enable media.webspeech.recognition.force_enable

(remember to turn em off after)

And then try official MDN example for Speech Recognition: https://mdn.github.io/web-speech-api/speech-color-changer/

Sorry for the trouble!

It turned out to be an issue in my code! Again, sorry for that and thanks for reporting this!

Should work now on FF, with those flags enabled (tested on FF78, MacOS)

a video proof that it does work in chrome: https://twitter.com/kddsky/status/1283764775151960071 :)