I’ll keep this short. If you’re looking to use the speech recognition capabilities of the Web Speech API on safari mobile (or safari for that matter) 7.1+ forget it. If you check out caniuse.com you’ll see that mobile safari only supports the speech synthesis part of the API (text –> speech). Android on the part support speech recognition in the browser. Just make sure to use the prefix “webkit” when calling the variable so:
should work for you.