Dual AI Chat
An advanced chat application demonstrates a unique dialogue paradigm: a user's query is first debated and refined by two different artificial intelligence roles before providing a final comprehensive answer. The project leverages the Google Gemini API to drive a logical AI (Cognito) and a skeptical AI (Muse), which work together to generate more robust, accurate and vetted responses.
Core Features - 🤖Dual AI Debate System: User input triggers internal discussions between Cognito (logical) and Muse (skeptical). This dialectical process aims to reduce AI illusions, explore multiple perspectives, and stress test information before giving a final answer. - 🔌Multi-API Backend Support: - Native support for the Google Gemini API. - Supports any OpenAI-compatible API, making it easy to interface with local models (such as Olama, LM Studio) or other hosted services. - 📝Shared Notepad with Undo/Redo: A collaboration space used by two AIs to record key points or draft solutions. Notepad content will be included in subsequent AI prompts, achieving status maintenance. Support full Markdown preview and multi-step ** undo/redo **. - ˇMultimodal Input: Users can upload pictures and text. AI can understand and discuss image content. - ˇHighly Configurable: - Model Selection: Switch between multiple Gemini models instantly. - API Configuration: Easily switch and configure Gemini or OpenAI-compatible APIs in in-app settings. - Discussion Modes: Supports "fixed-turn" conversations, or an "AI-driven" model where AI decides when to end the discussion. - 🔁Robust Error Handling: Contains automatic retry logic for API calls. More importantly, if the automatic retry fails, the app provides a ** Manual retry ** button to restore the entire conversation context from the failure point.
