If we could get a HIPAA compliant version that uses Groq's Whisper v3 Large for speech recognition and one of the smarter models for formatting/cleaning up (eg Sonnet 3.5 or GPT4o)... you will have just created the absolute greatest medical dictation app in existence, far better than Dragon for medical professionals. Hospitals routinely pay $1000 a year for Dragon per user.

I know it might be super onerous to get HIPAA compliance and maybe its a project not worth the effort, but you do that and man I am going to be evangelizing the shit out of this to my medical colleagues.

I think groq, Anthropic, OpenAI etc all offer HIPAA compliant versions of their stuff.

Note: my understanding is there is nothing wrong currently with using even the cloud providers on SuperWhisper so long as you dont include specific personally identifying information (eg full name, exact dates, locations smaller than a state per HIPAA rules) but it would just be less onerous if the user didn't have to worry about not including these in their dictations.