Web Speech API vs Google Speech API

Question

I've been using MDN Speech APIs (using things like window.speechSynthesis and new SpeechSynthesisUtterance). Firstly, I thought this API was the same as Google's Speech API, but I guess it is not?

Secondly, it's been working great, but there are some challenges I've been facing:

• only available in Chrome and Safari
• voices do seem robotic sometimes
• transcribed texts are awkward with numerals

Firstly, I'm curious if Google's Speech API solves these issues. Whas your experience been with Google's Speech API?

• Does the API work across any device? I want to build a game in Unity with Speech. Can I use it within Unity so long as I'm querying the API?

• Is it better at recognizing contexts?

  • spoken si yo hablo espanol -> Does not capture si
  • spoken me llamo Dan -> does not capture proper noun Dan

• Does it have less robotic voices?

Thank you!

---

Note: I am familiar with the use cases for both APIs, I'm looking more for the performance in each of the areas listed above.