Codex CLI

Add the apixly gateway as a custom model provider in Codex's config.toml, using the OpenAI-compatible chat wire API.

config.toml

toml

# ~/.codex/config.toml
model = "gpt-5.5"
model_provider = "gateway"

[model_providers.gateway]
name = "gateway"
base_url = "https://apixly.ai/v1"
env_key = "GATEWAY_API_KEY"
wire_api = "chat"

Run

Export the key under the env_key name you configured, then start Codex:

bash

export GATEWAY_API_KEY="YOUR_API_KEY"
codex

Use wire_api = "chat" — the gateway speaks OpenAI chat/completions. The key must be bound to an OpenAI group. Config keys can move between Codex versions; check codex --help if a field is rejected.