Is your feature request related to a problem?
Bad transcription on technical words and function names etc.
Describe the solution you'd like
FluidVoice does an on device screen read, parses text, potentially does OCR and feeds that as context to the recognizer model to produce the right transcription.
Describe alternatives you've considered
Adding to dictionary and instant rules but it is impossible to do it for custom names etc that appear on screen
Additional context
Apparently, VoiceInk does this (proprietary)
Is your feature request related to a problem?
Bad transcription on technical words and function names etc.
Describe the solution you'd like
FluidVoice does an on device screen read, parses text, potentially does OCR and feeds that as context to the recognizer model to produce the right transcription.
Describe alternatives you've considered
Adding to dictionary and instant rules but it is impossible to do it for custom names etc that appear on screen
Additional context
Apparently, VoiceInk does this (proprietary)