Codex CLI

Mac System Configuration

1. Manual Configuration

Required Downloads

For M-series chips, click to download: https://nodejs.org/dist/v24.13.1/node-v24.13.1.pkg

For Intel chips, click to download: https://nodejs.org/dist/v24.13.1/node-v24.13.1.pkg

• Note: Click Next throughout the installation. Do not change the installation path!

Installing and Configuring Codex

(1) Run the following commands separately in Terminal ⬇️ (Press Control + Space, type Terminal, then press Enter.)

# Enter in Terminal
sudo npm install -g @openai/codex

# If the download fails due to network issues, use the following command (turn off your VPN/proxy)
sudo npm install -g @openai/codex --registry https://registry.npmmirror.com

# Command to verify the installation
codex --version

• Password means your computer password. Type it and press Enter (the characters will not be displayed).

(2) Run vim ~/.codex/config.toml in Terminal, then enter the following file contents ⬇️

model_provider = "codex"
model = "gpt-5.6-sol"
model_reasoning_effort = "high"
disable_response_storage = true


[model_providers.codex]
name = "codex"
base_url = "https://api.openoxai.com/v1"
wire_api = "responses"
requires_openai_auth = true

(3) Run vim ~/.codex/auth.json in Terminal, then enter the following file contents ⬇️ (sk-xxx is the API key you created.)

{
    "OPENAI_API_KEY": "sk-xxx"
}

(4) Finally, run codex in Terminal to start using it:

Enter /model to choose an available model. Use the exact model ID from the live model catalog, for example gpt-5.6-sol.

Modifying the Configuration File (Method 2) (Use This for IDEs)

Install CC Switch

Mac installer

IDE Extension Setup (Configure Only If Needed; Otherwise, Skip This Step)

⬇️ Integration: After completing the configuration above and confirming that responses work normally, open a new VS Code / Cursor / Trae window and download the Codex extension to start using it.