OpenCodex serves as a local gateway for Codex Desktop, enabling third-party API integration and enhanced functionality. It targets users seeking streamlined control over AI models and desktop interactions, offering a plug-and-play solution with a web dashboard, custom computer use engine, and a Vision Bridge for text-only models. The primary benefit is simplified setup and expanded capabilities for AI workflows on macOS.

How It Works

This project acts as a local server that automatically configures Codex Desktop. It features a web dashboard for managing API keys, custom models, and logs. The core "Computer Use" engine leverages macOS native APIs for direct mouse, keyboard, and window control, plus screenshot capabilities. An adaptive router optimizes CLI-based browser control, achieving over 60% speed improvements. The "Vision Bridge" allows text-only models to process visual input by compressing screenshots and generating descriptions via a multimodal model.

Quick Start & Requirements

• Prerequisites: macOS, Node.js v18+, Codex Desktop installed.
• Install & Run:

  git clone https://github.com/AITabby/opencodex.git
  cd opencodex
  npm install
  npm start
  

  Server starts, opening http://localhost:8765/dashboard.
• Links: Demo Video (English) [YouTube link implied].

Highlighted Details

• Zero-Config Setup: Auto-patches Codex Desktop configuration (~/.codex/config.toml).
• Web Dashboard: Bilingual (EN/中文) interface for graphical management of API keys, endpoints, and custom models. Includes live SSE logs and one-click Codex restart/reset.
• Voice Integration: Direct control panel settings for Groq STT (Whisper) and Doubao TTS (Volcengine V3 API).
• Computer Use & Adaptive Routing: Enables native macOS mouse/keyboard/window control and screenshot capture. Adaptive routing significantly speeds up CLI browser control.
• Vision Bridge: Allows text-only models to interpret images by compressing screenshots, generating descriptions, and injecting them into prompts. Supports OpenAI-compatible vision models.

Maintenance & Community

The README provides no specific details on maintainers, community channels (e.g., Discord, Slack), sponsorships, or a public roadmap.

Licensing & Compatibility

The README does not specify a software license. This omission requires clarification regarding usage rights, particularly for commercial applications or integration into closed-source projects.

Limitations & Caveats

The project is explicitly designed for macOS only and requires Codex Desktop installation. The absence of a stated software license is a significant caveat for adoption.