How to implement local speech recognition exclusively on WatchOS? I think there was mention of local ASR, this year?
Use case is for a visually impaired person requesting information by raising their wrist and speaking a small number of words for a variety of intents. Because they would be walking with a white cane, the UI must be hands free. This works for sighted folks, as well. The number of intents are too diverse for creating individual Siri shortcuts.
Ideally, the Watch App would call something akin to SFSpeechAudioBufferRecognitionRequest. Is this available? Where can I find the documentation?