I'm excited about local AI because of the privacy characteristics (you can run FreeChat without even connecting to the internet). I think that to amplify our abilities to the fullest extent we'll want information to flow fluidly back and forth between our brain and the AI. I'm pretty wary of what I'd plug my uncensored brain into, but keeping everything on-device feels most trustworthy to me. I was blown away by the llama.cpp demos and want to 1) make something useful for myself and 2) push more people towards open models and local AI.

This is my first foray into packaging a local AI product for less technical users. The UX is basically cloned from other chat apps out there.

Please join the TestFlight if you're a mac user interested in local AI and want to see how many tokens/second you can get on your machine.