MackingJAI provides a local proxy to access Apple Intelligence and ChatGPT desktop applications via OpenAI or Ollama compatible APIs. This tool is designed for developers and users who want to integrate these powerful AI models into their existing workflows without direct API key management or complex setup, effectively mocking API calls to local applications.

How It Works

MackingJAI acts as a local server that intercepts API requests intended for OpenAI or Ollama. It then forwards these requests to either the Apple Intelligence models running on macOS or the ChatGPT desktop application. This approach bypasses the need for API keys and allows any application designed to work with OpenAI or Ollama APIs to seamlessly interact with local AI models.

Quick Start & Requirements

• Installation: Download and install the DMG file from releases. Click "Install Shortcut" from the menu icon.
• Prerequisites:
  • macOS Sonoma 14.5 or later with Apple Intelligence enabled for Apple models.
  • ChatGPT Desktop app installed and running for ChatGPT models.
• Configuration: Set API base to http://127.0.0.1:11435/v1/ for OpenAI API compatibility, or change Ollama's base URL to http://127.0.0.1:11435. Permissions may be required for API call interception.
• Documentation: README

Highlighted Details

• Supports both on-device and cloud-based Apple Intelligence models (apple_local, apple_cloud).
• Compatible with various ChatGPT models including GPT-4o, GPT-4.5, and older versions.
• Integrates with popular tools like LangChain, Raycast AI, and GitHub Copilot.
• No API key is required; any placeholder key will suffice.

Maintenance & Community

The project is maintained by 0ssamaak0. Further community engagement channels are not explicitly listed in the README.

Licensing & Compatibility

The repository does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

Functionality is limited by the underlying Apple Intelligence or ChatGPT desktop application's capabilities, including available models, rate limits, and generation speed. Parameters like temperature and top_p are not supported, and image input is not possible.