Using TOKENOPEN API with Codex CLI

Use OpenAI Codex CLI with TOKENOPEN API to access platform models directly from your terminal

Codex CLI is OpenAI's official open-source command-line coding agent. It can read and write files, execute shell commands, and complete complex programming tasks directly in your terminal. By pointing it at the TOKENOPEN API endpoint, you can seamlessly use all models available on the platform.

Step 1: Install Codex CLI

Prerequisite: Node.js 22+

# Add NodeSource repository
sudo curl -fsSL https://deb.nodesource.com/setup_lts.x | sudo -E bash -

# Install Node.js
sudo apt-get install -y nodejs

Verify the Node.js installation:

node --version
npm --version

Install Codex CLI:

npm install -g @openai/codex

Verify the installation:

codex --version

Option 1: Homebrew (recommended)

brew install --cask codex

Option 2: npm

npm install -g @openai/codex

Verify the installation:

codex --version

Run in PowerShell:

powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"

Or install via npm:

npm install -g @openai/codex

Verify the installation:

codex --version

Codex CLI supports macOS, Linux, and Windows (native or WSL2). Make sure Node.js 22+ is installed before proceeding.

Step 2: Configure TOKENOPEN API

Codex CLI uses ~/.codex/config.toml to manage the API connection.

Get an API Key

Configure TOKENOPEN API

Create the config directory and open ~/.codex/config.toml:

mkdir -p ~/.codex
vim ~/.codex/config.toml

Add the following content:

model_provider = "proxy"
model = "gpt-4o"
disable_response_storage = true

[model_providers.proxy]
name = "proxy"
base_url = "https://www.tokenopen.ai/v1"
wire_api = "responses"
requires_openai_auth = false
env_key = "OPENAI_API_KEY"

Press Esc, then type :wq! to save and quit.

Set the environment variable for the current session:

export OPENAI_API_KEY="replace-with-your-api-key"

To persist across sessions, add it to ~/.bashrc (bash) or ~/.profile:

echo 'export OPENAI_API_KEY="replace-with-your-api-key"' >> ~/.bashrc
source ~/.bashrc

Create the config directory and open ~/.codex/config.toml:

mkdir -p ~/.codex
vim ~/.codex/config.toml

Add the following content:

model_provider = "proxy"
model = "gpt-4o"
disable_response_storage = true

[model_providers.proxy]
name = "proxy"
base_url = "https://www.tokenopen.ai/v1"
wire_api = "responses"
requires_openai_auth = false
env_key = "OPENAI_API_KEY"

Press Esc, then type :wq! to save and quit. Alternatively, open it in TextEdit:

open -e ~/.codex/config.toml

Set the environment variable for the current session:

export OPENAI_API_KEY="replace-with-your-api-key"

To persist across sessions, add it to ~/.zshrc (default shell on macOS):

echo 'export OPENAI_API_KEY="replace-with-your-api-key"' >> ~/.zshrc
source ~/.zshrc

Run the following commands in PowerShell to create the config file automatically:

New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.codex" | Out-Null

@'
model_provider = "proxy"
model = "gpt-4o"
disable_response_storage = true

[model_providers.proxy]
name = "proxy"
base_url = "https://www.tokenopen.ai/v1"
wire_api = "responses"
requires_openai_auth = false
env_key = "OPENAI_API_KEY"
'@ | Set-Content -Path "$env:USERPROFILE\.codex\config.toml" -Encoding UTF8

Set the environment variable for the current session:

$env:OPENAI_API_KEY = "replace-with-your-api-key"

To persist permanently (user environment variable):

[System.Environment]::SetEnvironmentVariable("OPENAI_API_KEY", "replace-with-your-api-key", "User")

Or open the config file manually in Notepad:

notepad "$env:USERPROFILE\.codex\config.toml"

Step 3: Use Codex CLI

Launch Codex from inside your project directory:

codex

Launch Codex CLI

First-time Setup

On the first run, Codex will ask whether to trust the current directory:

First launch

Option 1: Allow Codex to execute in the current directory
Option 2: Exit

Switch Models

Inside the interactive interface, type /model to switch models:

Switch model

You will then be prompted to select a reasoning level:

Select reasoning level

Supported GPT and other models on TOKENOPEN (examples):

Model ID	Description
`gpt-5.5`	GPT-5.5, latest flagship model
`gpt-5.4`	GPT-5.4, balanced performance & cost
`gpt-5.4-mini`	GPT-5.4 Mini, low-latency lightweight model
`gpt-5.4-nano`	GPT-5.4 Nano, ultra-low-cost fast model

For the full model list and pricing, visit the Model Marketplace.

Start Chatting

Once the model is set, type your task or question in the prompt and press Enter:

Start using

Common Commands

Command	Description
`/model`	Switch model and reasoning level
`/init`	Create an AGENTS.md config file
`/status`	View current session config and token usage
`/approvals`	Configure which actions require approval
`/review`	Review and find issues in code changes

Non-interactive Mode

To run a one-shot task without entering the interactive interface:

# Run a single task
codex exec "Generate a Chinese version of the README.md in the current directory"

# Non-interactive mode
codex -p "Write a bubble sort algorithm in Python"

Codex CLI can read and modify local files and execute terminal commands, making it ideal for code review, refactoring, and generating tests. Using it inside a Git repository is recommended so you can track changes.

Using TOKENOPEN API with Codex CLI

On this page