OSX's speech assistant must be the most underrated feature of any product ever.

The complexity of speech recognition algorithms is colossal. I have been tracking their progress for years, as I have injuries from using the computer and depend upon speech recognition.

What is more, the recognition accuracy seems to be at least as good as Dragon/Nuance, without the appalling buggy front-end. I don't even require a custom microphone any longer. The MacBook's inbuilt microphone seems to be enough.

So I cannot say I am not grateful.

However, some things still drive me up the wall. If I say 'on' it will usually interpret 'en'. Which isn't even in the English language. And if I say 'script', it inserts 'prescription'. Which makes me want to punch something. If I want to say 'prescription' I will say 'prescription'.

Is there some way to disable these idiot presumptions? Some kind of "post translation" scripting I could do? Some way of intercepting the translated text and modifying it before it reaches the target app?


Many times, alternative words are available, as shown by a blue underline. Right-clicking the word and selecting the alternative word will 'train' the system over time to better understand what you are saying. Also make sure that Enhanced Dictation is enabled in System Preferences.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .