Want to transcribe recordings, any ideas where to start?

[reg98643t]

Essentially my team includes a few people with accessibility issues that necessitates they take a lot of notes in order to keep across what was discussed. We want a way to transcribe our meetings in a way that doesn't involve sending confidential information to a third party. We've got the recording setup fine, so my job now is trying to get some kind of transcript that can separate different speakers etc.

I've got LocalAI running well on my server, it can transcribe speech if I speak directly to it, but I've kind of hit a wall in where to go from here in running it as a post-process and if there's a model that specifically can split out different speakers.

Can anyone suggest where I could start getting it working to this purpose?