Codex Integration

This guide shows how to configure Codex CLI to use model capabilities through the unified R9S API.

Prerequisites

Before you start, make sure you have:

1. Codex CLI installed

   Follow the official Codex CLI installation instructions to install Codex locally.

2. An R9S API key

   Sign in to the R9S platform, create an API key, and copy it.

   Store this key securely. Codex will read it from an environment variable.

Configure Codex CLI

Edit your Codex configuration file:

vim ~/.codex/config.toml

Add the following configuration:

model_provider = "r9s"
model_reasoning_effort = "high"
model = "gpt-5.3-codex"
web_search = "disabled"

[model_providers.r9s]
name = "R9S"
base_url = "https://api.r9s.ai/v1"
env_key = "R9S_API_KEY"

Configuration notes:

• model_provider must match the provider name under [model_providers.r9s].
• env_key is the environment variable name that stores your API key. Do not put the API key itself in config.toml.
• To switch models, update the model value in config.toml.

Set the API Key Environment Variable

For the current terminal session, set:

export R9S_API_KEY="YOUR_R9S_API_KEY"

Replace YOUR_R9S_API_KEY with your actual R9S API key.

For long-term use, add the export command to your shell configuration file, such as ~/.zshrc, ~/.bashrc, or ~/.config/fish/config.fish.

Set the environment variable before starting Codex. Codex reads the API key from the variable specified by `env_key`.

Start Codex

Open your project directory and start Codex:

cd /path/to/your/project
codex

If the configuration is correct, Codex will send model requests through the R9S API.