Thanks Lukas :D

[lenaxia]

Hey Lukas, just wanna say this is a great app. This was exactly what I was looking for but didn't have the skills to build. I've been using it for several days now and many meetings, and its absolutely spot on, so thanks for putting this together, and for making it really feature rich for something so very niche.

I will likely open a few more PRs down the road to add some additional functionality.

Some of the things I'm thinking about (in more or less this order):

• Enable remote whisper server as onboard whisper processing kills battery life and makes my laptop run hot
• Improved diarization to support more than "me" and "they", as many of my calls have multiple speakers in "they"
• Add a mute button for my own mic (my kid sometimes is screaming in the background)
• Automatic time stamps in the notes field (and also have contents survive tab switches b/w history and post)
• Allow for modifying the prompts (maybe, this one is a bigger lift and not sure if I want it enough)

Let me know if any of these don't align with where you see the app going.

[lukasbach]

Hey lenaxia, thanks for your kind words, happy that the app helps you! I don't have any specific plans for next steps with Pensieve, right now I'm focusing on some other topics, so nothing that would collide with what you had in mind. Feel free to work on the things you mentioned.

One note regarding diarization - Yes, I also figured that the current categorization is annoying with more than 2 people, but right now it's also pretty trivial, it just seperates the speaker based on from which audio stream it comes from (mic vs speaker output, "stereo audio diarization), and uses that feature from whisper.cpp. whisper.cpp also supports using tinydiarize for proper diarization, but I didn't got great results from using that, which is why I didn't spend a lot of time trying to integrate that. But maybe you will have a better experience with that.

Let me know if you have any questions while working :)