A modern, local AI chat application built with SwiftUI, Swift Concurrency, and The Composable Architecture (TCA), leveraging Apple Foundation Models for efficient local LLM interactions on macOS and iOS.
recording.mp4
✅ Local LLM Chat using Apple Foundation Models
✅ Streaming responses with smooth TCA-based state management
✅ Adaptive, Apple Intelligence-inspired UI with dynamic gradients and glass effects
✅ Structured local database with sharing-grdb for chat history
✅ Cross-platform macOS and iOS support with efficient window management
✅ Minimal, clear architecture optimized for extension and experimentation
The project utilizes The Composable Architecture (TCA) for clear, testable, and modular state management. The following diagram illustrates the current reducer structure:
%%{ init : { "theme" : "default", "flowchart" : { "curve" : "monotoneY" }}}%%
graph LR
AppReducer ---> AppDelegateReducer
AppReducer ---> OnboardingReducer
HomeReducer ---> DialogsListReducer
HomeReducer -- optional --> MessageListReducer
MessageListReducer ---> MessageListInputReducer
AppDelegateReducer(AppDelegateReducer: 1)
DialogsListReducer(DialogsListReducer: 1)
MessageListInputReducer(MessageListInputReducer: 1)
MessageListReducer(MessageListReducer: 1)
OnboardingReducer(OnboardingReducer: 1)