- Added GPT-5 model support (with automatic non-streaming fallback)
  - Improved chat UI: User/Assistant avatar icons and redesigned message bubbles
  - Changed model selection to native Picker for better usability
  - Fixed macOS streaming response stuttering issue
  - Code refactoring and stability improvements

Featured model section

Update some functions & Fix Bugs

Support GPT5

Multi-platform LLM support (Ollama, LM Studio, Claude, OpenAI)

- **Enhanced Settings Panel**: Added Temperature, Top P, Top K parameter controls with intuitive sliders
- **Connection Testing**: Added server connection status checker in settings
- **Multi-format File Support**: Extended file support from images only to images, PDF documents, and text files
- **Improved File Handling**:
  - PDF files: Automatic text extraction and inclusion in prompts
  - Text files: Direct content reading and inclusion in prompts
  - Images: Maintained existing image processing functionality
- **UI Improvements**: Enhanced settings interface with parameter descriptions and real-time value display

- Add token speed
- Support multi line input

Improved communication speed
Added cancel function during conversation