IMO, the current state of voice recognition in general is absolute garbage for anyone who doesn't speak American English natively. Dragon for Windows is touted as having great voice recognition, but it often fails to understand 'page up' / 'page down' and other simple commands. The built in system to OS X is ok if dictation with mistakes that need correcting by hand (aka keyboard) is something one desires; I can't imagine it ever is. Interestingly enough, the mobile voice recognition features seem to work a little better, but systems like Siri and Google's assistant are completely useless for getting actual work done. The accuracy of all these systems is such shit, that I have turned them all off as they are not even useful for dictating a plain English message into slack, let alone driving a computer and doing productive programming and system administration work. On the other hand, I've seen native speakers do exactly that which is impressive and also, I admit, makes me jealous as I don't expect to have access to this technology for many years to come.
Can you join the slack linked at talonvoice.com or send an email to the address in my hn profile?
I'd probably start by asking for audio samples of you saying a couple of sentences and commands in english
Then I'd try them locally against a couple of speech recognition engines
Then talk about potential issues and ways to influence recognition accuracy
Then I can help you try out Talon for both dictation and command input
As an anecdote, I've heard from some of my non-english-native users that they need to avoid certain word sounds in commands (by changing the command trigger words) because their accent doesn't emphasize them enough for good accuracy.
