Codex CLI provider

Use a GPT-5.5 API gateway with Codex CLI custom providers.

Codex CLI can use a custom provider when the gateway exposes a compatible Responses route, model ID, base URL, and API key environment variable. Zora provides those pieces through the Zora GPT/Codex model IDs returned by GET /v1/models.

Codex setup docs Model IDs

Recommended Codex config

model_provider = "zora"
model = "gpt-5.5"

[model_providers.zora]
name = "Zora"
base_url = "https://api.zora.io.vn/v1"
env_key = "ZORA_API_KEY"
wire_api = "responses"

Smoke test

# macOS / Linux
export ZORA_API_KEY="YOUR_ZORA_API_KEY"
codex exec -m gpt-5.5 "Reply with exactly: ZORA_CODEX_OK"

# PowerShell
$env:ZORA_API_KEY = "YOUR_ZORA_API_KEY"
codex exec -m gpt-5.5 "Reply with exactly: ZORA_CODEX_OK"

Beta test model order

Start with gpt-5.5 for the first Codex smoke test.
Use exact Zora model IDs from GET /v1/models.
Use exact model IDs from GET /v1/models; do not append extra model suffixes.
Move to another Codex ID only after the smoke test works.

When this setup is useful

You want GPT-5.5 model routes in a coding agent workflow.
You need an OpenAI-compatible API key that works with local CLI tools.
You prefer request packages and RPM limits instead of direct provider billing.
You need VietQR/VNPay payment for a coding API workflow.

FAQ

What does Codex CLI need from a custom provider?

Codex CLI needs a provider name, base URL, environment variable for the API key, model ID, and the correct wire API. Zora uses the Responses wire API.

Can I keep the API key out of the config file?

Yes. Store the key in ZORA_API_KEY and point the Codex provider config to that environment variable.