As per the Web API article
SpeechRecognition and SpeechSynthesis supported by Chrome only and it uses the server-based recognition engine
Does Server-based recognition Engine uses the Google Speech API ?