Can Parley Notes tell speakers apart?
Yes. Every transcript is split into turns and each speaker gets a label: Speaker 1, Speaker 2 and so on. This runs on the phone.
Parley Notes assumes two speakers unless you tell it otherwise. We found that counting speakers automatically is unreliable with a phone lying on a table so set the expected number under Recording options when you know it. If the count came out wrong, open the transcript and pick the real number. The lines are re-sorted in a second or two.
Naming people
Tap a speaker label to rename that speaker everywhere in the note: the transcript, the participants, the decisions and the action items. Parley Notes suggests names it heard in the conversation, like "thanks, Sarah". One line attributed to the wrong person can be moved on its own and any line's words can be corrected by tapping it.
Renaming does not rewrite the summary on its own because that would spend minutes from your plan. Tap Re-summarize when you want the prose to catch up.
How it learns voices
Your voice comes first so in Settings under Your voice record four short samples in the situations the screen lists. From then on your own lines carry your name.
The people you talk to come next because Parley Notes can keep a voice when you name its speaker so the next conversation with the same person is labelled on its own.
Every voiceprint stays on your phone. None of them is sent to our servers or to anyone else, they do not follow your account to a new phone and Delete my voice in Settings removes them. See Voices stay on your phone.