Integrate Codex

Codex is an AI coding tool for software development tasks. HiModels exposes an OpenAI Responses-compatible endpoint for Codex, and clients that read the same Codex configuration directory can reuse the provider configuration.

1. Install Codex CLI

Install Codex CLI with npm and verify that the command works:

npm install -g @openai/codex
codex --version

2. Configure the HiModels Provider

Create an API Key from HiModels API Keys.

Set the key in the current terminal environment:

# macOS / Linux
export HIMODELS_API_KEY="<your_himodels_api_key>"

# Windows PowerShell
$env:HIMODELS_API_KEY="<your_himodels_api_key>"

Create or edit ~/.codex/config.toml and add the HiModels provider:

model = "gpt-5.5"
model_provider = "himodels"
preferred_auth_method = "apikey"
model_reasoning_effort = "high"

[model_providers.himodels]
name = "HiModels"
base_url = "https://api.himodels.ai/v1"
env_key = "HIMODELS_API_KEY"
wire_api = "responses"

On Windows, the file is usually located at %USERPROFILE%\.codex\config.toml. Never commit a real API Key to your repository.

3. Start Using Codex

Open your project directory and start Codex:

cd /path/to/my-project
codex

To temporarily switch models, pass a model ID available on HiModels:

codex --model gpt-5.5
codex --model gpt-5.4