OpenAI Codex API key setup
OpenAI's Codex coding agent — CLI (local), IDE (VS Code), Web (cloud async), Sites. GPT-5.5 with AGENTS.md, slash commands, approval modes, role plugins. ChatGPT plans add IDE and Web surfaces; rate limits scale by tier.
Get your key
- Open the OpenAI Codex site: openai.com/codex/
- Create an account and generate an API key.
- Paste it into the Dracon BYOK provider keys — stored encrypted (AES-256-GCM), used browser-direct, works in every Dracon tool.
Free tier
Yes. Codex CLI is free with a ChatGPT account (sign-in required). The Codex usage page or limit banner in the active CLI session is the canonical way to see remaining allowance and reset windows.
Rate limits: Codex usage: 1x tier (~20 messages/5hr, ChatGPT Plus) — base models. 5x tier (~100 messages/5hr, ChatGPT Pro) — all models. 20x tier (~200 messages/5hr, Business/Enterprise) — highest tier. Pro is higher-allowance, not unlimited. Windows reset on a 5-hour rolling basis. API-key usage is on a separate billing lane.
Connection details
- Env var
OPENAI_API_KEY- Model lineup
- codex-1, codex-mini, gpt-5-codex, gpt-5, gpt-4.1
Gotchas
Browser-direct chat is not available for this provider: CLI tool, not an API. Dracon provider keys still cover it for tools that run elsewhere (for example a LiteLLM chain from the Dracon Router builder).
Our verdicts on this kind of provider
Independent comparisons we keep current as pricing and free tiers shift. A guide is setup; a verdict is whether it's worth setting up.