GitHub - Lywald/portfolio-MixAI: Multi-Model AI Chat

3 min read Original article ↗

MixAI - Multi-Model AI Chat Application

MixAI is a desktop application that enables simultaneous interaction with multiple AI models (Claude, GPT-4, and Gemini) in a unified chat interface. It allows users to communicate with all models at once or target specific AI assistants for specialized responses.

image

image

Features

  • Multi-Model Integration: Seamlessly interact with Claude (Anthropic), GPT-4 (OpenAI), and Gemini (Google) in a single interface
  • Targeted Interactions: Use @all, @claude, @gpt, or @gemini to direct questions to specific AI models
  • Concurrent Processing: Asynchronous message processing for smooth user experience
  • Conversation History: Full context preservation across all AI models
  • Clean Qt-based GUI: Modern, responsive desktop interface

Prerequisites

  • Python 3.8 or higher
  • PyQt5
  • Active API keys for:
    • Anthropic (Claude)
    • OpenAI (GPT-4)
    • Google AI (Gemini)

Installation

  1. Clone the repository:
git clone https://github.com/Lywald/MixAI.git
cd MixAI
  1. Create and activate a virtual environment (recommended):
python -m venv venv
# On Windows
venv\Scripts\activate
# On macOS/Linux
source venv/bin/activate
  1. Install required packages:
pip install anthropic openai google-generativeai PyQt5

Configuration

Before running the application, make sure to set up your API keys in the lines 33/34/35 of mixai.py file:

self.claude = anthropic.Anthropic(api_key="your-anthropic-api-key")
self.gpt = OpenAI(api_key="your-openai-api-key")
genai.configure(api_key="your-google-api-key")

Usage

  1. Start the application:

  1. The chat interface will open automatically.

  2. To use the chat:

    • Type a message and press Enter or click Send to get responses from all AIs
    • Use @all, @claude, @gpt, or @gemini followed by your message to target a specific AI
    • Previous conversation context is maintained for all AIs

Example commands:

Hello, can you all introduce yourselves?
@claude What do you think about poetry?
@gpt Can you analyze this code snippet?
@gemini Please explain this scientific concept.

Architecture

The application is built with three main components:

  1. MixAI Class: Core logic for AI model integration and message processing
  2. AIWorker: Threaded worker for asynchronous message processing
  3. MixAIGUI: Qt-based graphical user interface

Key Features Implementation

  • Conversation History: Maintains context across all AI models
  • Asynchronous Processing: Prevents UI freezing during API calls
  • Error Handling: Robust error catching and user feedback
  • Model Targeting: Flexible message routing system

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

  • Anthropic for Claude API
  • OpenAI for GPT-4 API
  • Google for Gemini API
  • Qt team for PyQt5

Support

For support, please open an issue in the GitHub repository or contact the maintainers.

Disclaimer

This project is not officially affiliated with Anthropic, OpenAI, or Google. Please ensure you comply with each AI provider's terms of service and API usage guidelines.