Codex CLI

Point OpenAI's Codex CLI at LiteRouter through its Responses endpoint.

Codex CLI talks the OpenAI Responses API, which LiteRouter serves at /v1/responses for every model. You add LiteRouter as a custom model provider, then pick a model.

Configure

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

model = "glm-5.2"
model_provider = "literouter"

[model_providers.literouter]
name = "LiteRouter"
base_url = "https://api.literouter.com/v1"
env_key = "LITEROUTER_API_KEY"
wire_api = "responses"

Then export your key in the shell, using the same name as env_key above:

export LITEROUTER_API_KEY="your_literouter_api_key"

Run it

codex