explaingit

xqnode/codex-helper

110Rust
This is a quick first-pass explanation. The richer sections — use-cases, tech stack, setup, prompts — are still being generated.

TLDR

Codex Helper is a small system tray application for Windows and macOS, written in Rust, that lets you use OpenAI's Codex Desktop coding tool with Chinese AI models instead of OpenAI's own.

Mindmap

A visual breakdown will appear here once this repo is fully enriched.

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

In plain English

Codex Helper is a small system tray application for Windows and macOS, written in Rust, that lets you use OpenAI's Codex Desktop coding tool with Chinese AI models instead of OpenAI's own. Models you can switch to include DeepSeek, Tongyi Qianwen (from Alibaba), Moonshot, Zhipu GLM, and MiniMax, plus any OpenAI-compatible third-party endpoint. Switching happens from the tray icon menu, not through config files or a terminal. The tool works by running a local proxy on your machine at a fixed port. Codex is configured once to point at that proxy, and then Codex Helper forwards your requests to whichever model you have selected. It also handles the format difference between OpenAI's Responses API and the Chat Completions format that most Chinese providers use, so you do not have to think about that translation layer. Setup is meant to be minimal. On first launch the app detects whether Codex is already installed and auto-fills any existing API key it finds in your environment. A guided window appears if neither is present, with direct links to apply for keys from each supported provider. The settings panel lets you manage API keys (stored encrypted, never uploaded), set a startup preference, and back up or restore your configuration. A CLI mode mirrors every tray action for users who prefer the command line. The uninstaller restores your original Codex config from a backup and removes startup entries, leaving no leftovers. The application binary is under 10 MB and is expected to use less than 30 MB of memory at runtime. The project is primarily aimed at users who want cost-effective alternatives to OpenAI without editing configuration files or understanding how proxy routing works.

Open on GitHub → Explain another repo

← xqnode on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.