We are currently working on an app, that would greatly benefit from speech-to-text functionality.
We are using the latest version of MRTK and noticed, that while MRTK provides speech-to-text functions, it relies on Windows Speech to do so. Now Windows Speech is only available on Windows and UWP platforms (Hololens), but not on Android or any other plattform. Seemingly, Magic Leap is aware of this and implemented their own approach to enable the keyword recognition via their own methods.
The Dictation functionality is technically part of the same module, but currently still relying on Windows Speech. Does anyone - maybe directly at Magic Leap - know, if there are ongoing efforts to replace the Windows Speech part in the dictation functionality as well? Is there maybe already a ready-to-use equivalent, which we could rely on? Otherwise we'd probably try and go the cloud-based way of Azure Speech - but it would have been nice to stick with the already existing methods, if possible.
The Magic Leap SDK does not provide a speech dictation feature, instead it support Voice Intents, which allow your application to react to specific keywords or phrases such as "Change color".
You will need to use a third party solution for the dictation feature, at this time we do not have any recommendations but it looks like there are a few options on the Unity Asset Store.