Skip to main content

​
Overview

Codex CLI is OpenAI’s open-source command-line coding agent. It can read your codebase, propose changes, execute commands, and iterate based on feedback — all from the terminal. By pointing Codex CLI to FoxAPI, you can use it with any OpenAI-compatible model available through FoxAPI’s API.
Codex CLI supports a config.toml configuration file for defining custom model providers, making it easy to connect to FoxAPI without modifying environment variables for the base URL.

​
Prerequisites

FoxAPI API Key

An active FoxAPI API key with access to OpenAI-compatible models.

Node.js 22+

Codex CLI requires Node.js version 22 or later.

​
Installation & Configuration

1

Install Codex CLI

npm install -g @openai/codex
Verify the installation:
codex --version
2

Set Your API Key

Codex CLI reads the OPENAI_API_KEY environment variable for authentication. Set it to your FoxAPI API key:
Add the following line to your shell profile (~/.bashrc, ~/.zshrc, etc.):
export OPENAI_API_KEY="sk-your-foxapi-api-key"
Then reload your shell configuration:
source ~/.zshrc  # or source ~/.bashrc
3

Configure FoxAPI as a Model Provider

Create or edit the Codex CLI configuration file:
Edit ~/.codex/config.toml:
model = "gpt-4.1"
model_reasoning_effort = "medium"
model_provider = "foxapi"

[model_providers.foxapi]
name = "FoxAPI"
base_url = "https://api.foxapi.cc/v1"
env_key = "OPENAI_API_KEY"
wire_api = "responses"
The wire_api field must be set to "responses". The "chat" option is deprecated and may cause unexpected behavior.
Configuration fields explained:
FieldDescription
modelDefault model to use (can be overridden with --model flag)
model_reasoning_effortReasoning effort level: "low", "medium", or "high"
model_providerName of the custom provider (must match the [model_providers.xxx] section)
base_urlFoxAPI’s API endpoint
env_keyEnvironment variable name that holds your API key
wire_apiAPI protocol to use ("responses" is required)
4

Start Using Codex

Navigate to your project and launch Codex:
cd /path/to/your/project
codex
Or run a one-off command:
codex "Add input validation to the signup form"

​
Verify the Connection

Test the connection with a simple prompt: 
codex "Say hello and confirm you are working"
If the configuration is correct, Codex will respond through FoxAPI’s API.
ModelBest For
gpt-4.1Complex multi-step coding tasks (recommended default)
gpt-4oGeneral coding tasks with balanced speed and quality
o4-miniFast, cost-effective coding assistance
You can override the default model from config.toml when launching Codex: 
codex --model gpt-4o "Refactor the auth module"

​
Troubleshooting

  • Verify your OPENAI_API_KEY is a valid FoxAPI key.
  • Ensure the key has not expired and has sufficient balance.
  • If using config.toml, confirm the env_key field matches the environment variable name you set (default: OPENAI_API_KEY).
  • Check that base_url in config.toml is set to https://api.foxapi.cc/v1 (with /v1).
  • Verify your network can reach api.foxapi.cc.
  • Confirm the model name is correct and available on FoxAPI.
  • Check your FoxAPI account for model access permissions.
  • Confirm the file is located at ~/.codex/config.toml (macOS/Linux) or C:\Users\{username}\.codex\config.toml (Windows).
  • Validate the TOML syntax — ensure strings are properly quoted and section headers use square brackets.
  • Check that model_provider value matches the [model_providers.xxx] section name exactly.