Tools

Codex CLI

OpenAI's terminal coding agent, running on Omnizen models.

What you get

Codex CLI talks to the OpenAI Chat Completions API. Point it at Omnizen and you keep its tool surface (file edits, bash, plan mode) while we route to the cheapest credible frontier model behind the scenes.

Install

bash

npm install -g @openai/codex

Configure

Two environment variables in your shell profile:

~/.zshrc or ~/.bashrcbash

export OPENAI_BASE_URL="https://api.omnizen.ai/v1"
export OPENAI_API_KEY="sk-omni-…"

Or use Codex's own config file if you'd rather keep it scoped:

~/.codex/config.tomltoml

[provider]
base_url = "https://api.omnizen.ai/v1"
api_key  = "sk-omni-…"

Run it

bash

codex "add rate-limit headers to /v1/messages"

Pass --model to pin a specific Omnizen model — see /models for the full list.

Tips

Approval gates still work. Codex's interactive approval flow is unchanged — Omnizen sits behind it.
Use scoped keys from the dashboard if you want a per- repo spend cap. Codex picks up the env var per shell.