BasedHardware/omi ? reverse-engineered prompt

I want to build an AI assistant that acts like a "second brain". It should capture my conversations and whatever is on my computer screen, transcribing everything in real time. From that information, it should be able to automatically generate summaries and action items. The main feature is an AI chat that remembers everything I've seen and heard, so I can ask it questions about my past.

This should be a full system with apps for different devices. I'll need a desktop app for macOS, a mobile app that works on both iOS and Android (I think Flutter is a good choice for this), and a web app too. It also needs to be able to connect to wearable hardware like a smart necklace or glasses for capturing audio. All of these frontends should talk to a central backend that will handle the transcription, summaries, and AI chat features.