Get a key

Codex CLI

View as Markdown
On this page

Codex CLI works over OpenAI Responses and no longer supports Chat Completions. We answer over Responses as is: the response.* event stream, tools and reasoning.

Setup#

Add a provider to ~/.codex/config.toml:

~/.codex/config.toml
model = "gpt-6-astra"
model_provider = "aic"

[model_providers.aic]
name = "AI Creator"
base_url = "https://api.aicreator.app/v1"
env_key = "AIC_API_KEY"
wire_api = "responses"
stream_idle_timeout_ms = 300000

Codex reads the key from the environment variable named by env_key:

bash
export AIC_API_KEY=sk-aic-…
codex
  • base_url ends with /v1.
  • wire_api = "responses" — required.
  • stream_idle_timeout_ms — five minutes without events; on long runs we send keepalives, so the stream holds.

Models#

Set model to the model’s full id. Models tagged Codex on the pricing page work with Codex. The API itself gives the exact list for your key:

bash
curl "https://api.aicreator.app/v1/models?surface=responses" \
  -H "Authorization: Bearer $AIC_API_KEY"

A model that is not on the Responses protocol returns 400 model_not_available. Other common failures are in Troubleshooting.