I’m wondering if it would make sense to add an optional AI transcript enhancement feature. Ultimately the original transcript should still be accessible, but you could have an LLM analyze the transcript and make edits/revisions to effectively clean it up.
I have done this from transcripts generated by iOS voice memos, sent the transcript with a custom prompt to an LLM, and have gotten much better/more accurate transcripts as a result.
I have done the same thing. When I was using the Not-ta (I had to hyphenate as this board doesn’t like it) app, the transcriptions were awesome and live. It was pretty cool to watch the transcription live. I could even bookmark areas I needed to.
The only reason I went to a separate device was due to Voice Memo’s Not-ta, etc would stop recording if ANY thing happened on my phone such as a text, call, email, doorbell, etc