> For the complete documentation index, see [llms.txt](https://docs.world3.ai/world3/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.world3.ai/world3/how-to-guides/routerlink/tutorials/how-to-integrate-routerlink-in-codex-cli.md).

# How to integrate RouterLink in Codex CLI

{% hint style="info" %}
This documentation is provided for informational purposes only and demonstrates how to configure and use our API with third-party AI chat interfaces. Any third-party software, websites, or services mentioned are not operated, controlled, or endorsed by us.
{% endhint %}

## Overview

This tutorial guides you through integrating **RouterLink**—WORLD3's decentralized AI model routing infrastructure—with **Codex**, a personal AI assistant you run on your own devices.

### Purpose

RouterLink provides a unified API endpoint that routes requests to multiple AI model providers through a decentralized network. By integrating RouterLink with Codex, you gain access to a diverse range of large language models (LLMs) without managing multiple API keys or provider-specific configurations.

### What You Will Learn

By completing this tutorial, you will:

* **Configure Codex** to connect with RouterLink's API endpoint
* **Authenticate** using your RouterLink API credentials with custom headers
* **Add and manage AI models** available through the RouterLink network
* **Execute inference requests** through Codex's conversational interface

### Prerequisites

Before proceeding, ensure you have:

* A RouterLink API key (obtain one from the [Quick Start Guide](https://docs.world3.ai/world3/how-to-guides/routerlink/quick-start-guide))
* A web browser with internet access
* Basic familiarity with API configuration concepts

## Introduction to Codex

<figure><img src="/files/oxY7XibCDk6q6d91wzOW" alt=""><figcaption></figcaption></figure>

Codex CLI is OpenAI’s coding agent that you can run locally from your terminal. It can read, change, and run code on your machine in the selected directory. It’s [open source](https://github.com/openai/codex) and built in Rust for speed and efficiency.**Codex** Key Features:

* Terminal-based AI coding assistant
* Natural language → code generation
* Local execution with privacy
* Multiple automation modes
* Multi-language support
* Debugging, refactoring & testing automation
* Interactive workflow in CLI

### Step 1: Install Codex

Install [the newest LTS version of Node.js](https://nodejs.org/en/download/)

After installation is complete, open the terminal (Windows users use PowerShell) and check the version number.

* macOS, Linux, Windows:

```bash
npm i -g @openai/codex
```

Verify the installation.

```
codex --version
```

If the version number is displayed correctly, the installation is successful.

### Step 2: Configure Codex

Codex uses the `config.toml` file to parse provider parameters. Please locate and create/edit this file according to your operating system.

Edit or create `config.toml` file&#x20;

On macOS & Linux: \~/.codex/config.toml&#x20;

On Windows: C:\Users\\\<USERNAME>\\.codex\config.toml

Delete the previous content and replace it with the following.Configuration file:

```toml
model_provider = "routerlink"
model = "world3-router-north-america/openai/gpt-5.3-codex"

[model_providers.routerlink]
name = "RouterLink"
base_url = "https://router-link.world3.ai/api/v1"
env_key = "ROUTERLINK_API_KEY"
```

You can replace the "model" field with any model that is compatible with the OpenAI format.

### Step 3: Codex Loading and Validation

Open the terminal (Windows users use PowerShell). Replace **${API\_KEY}** below with your key (obtain one from the [Quick Start Guide](https://docs.world3.ai/world3/how-to-guides/routerlink/quick-start-guide)).

* macOS, Linux,

```toml
export ROUTERLINK_API_KEY="${API_KEY}"
```

* Windows PowerShell:

```toml
$env:ROUTERLINK_API_KEY = "${API_KEY}"
```

Launch codex

```toml
codex
```

Select "Yes" to trust this folder.

<figure><img src="/files/vXyLk0aaoZEOK98OBUJJ" alt=""><figcaption></figcaption></figure>

Now you can use Codex.