“AI” text prediction runs locally. Microphone is for voice to text functionality.
As for the keyboard itself. Ehhhhh. It’s lacking UX features to make it actually usable. I dailied it for a month and had far more typos, text prediction broke whenever a number or symbol was fat fingered into the string. Finding symbols you need was worse than gboard & SwiftKey.
I really want there to be a great open-source keyboard, but none actually deliver on UX atm.
What I find hilarious is all these companies doing this shit after all the advancements in programming languages and paradigms in the last few years.
Thanks to tools like Node.js, React, Flask, Reflex, OpenAPI Gen, GoLang, and more, people that are fed-up and have the know-how can stand up competing technology in record time.
I look forward to see what comes out of this corporate power grab. Hopefully there’s not a lot of pain and suffering alkng the way.