How techjarves/uncensored-local-ai-multiplatform works
Uncensored Local AI is a Flutter app for running open-source language models on-device. The main user journey is to obtain a model through the model library, load it, chat locally, and retain conversations in local history. A separate settings-driven capability exposes an OpenAI-compatible local HTTP API for external clients. The README also documents model import, loading progress, cancellation, and live generation metrics. Flutter implementation files are present in the tree, but their source was not sampled; relationships below rely on README statements, so internal call ownership cannot be verified.
Main entry points: home_screen.dart Home screen · model_library_screen.dart Model library · local_api_server_service.dart Local API server · settings_screen.dart Settings screen
Components 20
App experience
- Home screen
lib/screens/home_screen.dart - Settings screen
lib/screens/settings_screen.dart - Logs screen
lib/screens/logs_screen.dart - Log service
lib/services/log_service.dart
Model runtime
- Model library
lib/screens/model_library_screen.dart - Model card
lib/widgets/model_card.dart - Model controller
lib/controllers/model_controller.dart - Model manager
lib/services/model_manager.dart - Model catalog
assets/models_catalog.json - LLM service
lib/services/llm_service.dart - Loading metrics
lib/models/download_state.dart
Conversation
- Chat controller
lib/controllers/chat_controller.dart - Chat UI
lib/screens/home_screen.dart - Chat history
lib/widgets/chat_sidebar.dart - Chat storage
lib/services/chat_storage_service.dart - Local history store
lib/services/chat_storage_service.dart
Local integration
- Local API server
lib/services/local_api_server_service.dart - API settings
lib/screens/api_endpoints_screen.dart
Other components
- User
- API client
Connections 15
- User → Model library: choose model
README.md - Model library → Model catalog: browse catalog
README.md - Model library → Model manager: download or import
README.md - User → Home screen: start chatting
README.md - Home screen → LLM service: chat locally
README.md - Home screen → Chat history: view history
README.md - Chat history → Chat storage: load history
README.md - Chat storage → Local history store: persist locally
README.md - User → Settings screen: configure server
README.md - Settings screen → Local API server: toggle server
README.md - API client → Local API server: send API requests
README.md - Local API server → LLM service: serve completions
README.md - Model manager → Loading metrics: report progress
README.md - User → Logs screen: inspect logs
- Logs screen → Log service: read logs