How it works: This app tries to run AI inference locally on your device for privacy and speed. While the model downloads, it uses a remote API. Once ready, responses are generated entirely on your device!