Codex CLI

Quick checklist

1. Install or update Codex CLI.
2. Edit ~/.codex/config.toml and keep provider settings in the user-level file.
3. Create model_providers.lingai with base_url https://api.lingapi.ai/v1, wire_api = "responses", requires_openai_auth = true, and supports_websockets = false.
4. Create ~/.codex/auth.json with OPENAI_API_KEY set to your Ling.AI key, then run a small prompt.

Provider entry

Codex CLI reads provider profiles from ~/.codex/config.toml. Keep the API key in ~/.codex/auth.json and keep any fixed model id aligned with /v1/models.

Wire API and websocket

Use wire_api = "responses" for Codex-compatible models and set supports_websockets = false so this provider uses normal HTTP Responses requests.

CC Switch

Codex can also be configured from the Codex tool page in CC Switch. Use an API-key provider, base URL https://api.lingapi.ai/v1, and restart Codex CLI after switching.

Example

model_provider = "lingai"
model_reasoning_effort = "xhigh"
disable_response_storage = true
model_context_window = 1000000
model_auto_compact_token_limit = 900000
# model = "<model id from /v1/models>"

[model_providers.lingai]
name = "Ling.AI"
wire_api = "responses"
base_url = "https://api.lingapi.ai/v1"
requires_openai_auth = true
supports_websockets = false
# File: ~/.codex/config.toml

{ "OPENAI_API_KEY": "<YOUR_API_KEY>" }
# File: ~/.codex/auth.json