Open to test - MacOS & Windows - v0.0.2

New Features

Deep Thinking Mode

• Toggle between Normal and Deep Thinking modes for AI responses
• Deep Thinking mode uses higher temperature (0.9) and increased token limit (4096) for more thorough reasoning
• Perfect for complex problems requiring detailed analysis and creative solutions
• Easy-to-use toggle switch with brain icon next to model selector

RAG (Retrieval-Augmented Generation) Support

• Chatbot now maintains full conversation context across messages
• AI can reference previous questions and answers in the conversation
• Improved follow-up question handling with contextual awareness
• Seamless conversation flow with memory of entire chat history

Copy Message Functionality

• One-click copy button for all AI responses
• Material-UI icons with visual feedback (checkmark confirmation)
• Hover tooltips for better UX
• Automatically copies full message content to clipboard

Material-UI Icon Integration

• Upgraded from emoji to professional Material-UI icons
• Detailed prompt suggestions with clear placeholders
• Enhanced visual consistency across the application
• Icons: Lightbulb, Code, Summarize, Tips & Updates, School, Create

Improvements

• Prompt Suggestions: More detailed and actionable prompt templates with bracketed placeholders
• UI Polish: Better icon styling and hover effects
• Type Safety: Enhanced TypeScript definitions for Electron API
• Code Quality: Improved component structure and separation of concerns

Technical Updates

• Added @mui/material, @emotion/react, @emotion/styled dependencies
• Added @mui/icons-material for professional icon set
• Implemented Ollama chat API for conversation context support
• Enhanced LLM client with message history support

Bug Fixes

• Fixed TypeScript type definitions for Ollama chat method
• Improved error handling for LLM API calls
• Better conversation history management

Breaking Changes

None - fully backward compatible with v0.0.1