Want to try Diane - prototype for Voice to Whisper to GPT3 to Obsidian?

During the past week, Carl Heath and I have been publicly tinkering with a prototype of a kind of voice-recording app that we call Diane. Diane records your voice, or multiple voices, and automatically sends the audio to the Whisper service for transcription. After that you choose a template that instructs GPT3 how to summarize the transcription, and finally your result is saved in Obsidian.

I haven’t had time to continue working on this prototype, but many people have wanted to try Diane, so for those of you who have a bit of knowledge about how to run Python code on your computer, there’s code and instructions at github.com/mickekring/Diane. If you don’t know how to run Python code on your computer but still want to try, I have a short tutorial on how to do that (if you run Windows) here mickekring.se/sa-har-kor-du-ett-pythonscript-pa-din-windowsdator/

Things to know to run the script

• You need an API key from OpenAI
• You need to have Obsidian installed obsidian.md
• I’m not a programmer, so don’t blame me for bad code
• Only tested on Mac. But ChatGPT says the code works on Windows too. :)

How it works

Here you can watch the video about how I built the app and also see how it works.

History

Here you can read more about how Carl and I tinkered: mickekring.se/diane-en-prototypapp-for-rostanteckningar-bearbetat-av-en-massa-ai/

Questions?

As usual, just get in touch via any medium :)