OpenAI Codex CLI

Skip it — one-line setup

The script installs OpenAI Codex CLI from its official source, writes the gateway settings into the tool's own config file (backing up any existing one) and runs a live check. It asks for your gateway key, or take it from --key / --key-file. --uninstall restores exactly what it wrote.

curl -fsSL https://quantoport.com/agents/codex/setup | bash

1 · Install

Install OpenAI Codex CLI from its own official source.

npm install -g @openai/codex
# or Homebrew:
brew install --cask codex
# or the official script:
curl -fsSL https://chatgpt.com/codex/install.sh | sh

2 · Configure — point it at QuantoPort

Gateway base URL https://quantoport.com; get your key from the console.

International site https://quantoport.com (native /v1/responses):
1) Edit ~/.codex/config.toml:
   model = "gpt-5.6-sol"
   model_provider = "quantoport"

   [model_providers.quantoport]
   name = "QuantoPort"
   base_url = "https://quantoport.com/v1"
   env_key = "QUANTOPORT_API_KEY"
   wire_api = "responses"
2) Export the key: export QUANTOPORT_API_KEY=<your gateway key>
3) Run codex.
Note: Codex 0.153+ only accepts wire_api = "responses" (the old "chat" value errors out). The CN site https://api.shinlinkai.com serves only /v1/chat/completions, not /v1/responses, so use the international site for Codex.

3 · Verify

One request proves you are on our channel.

codex exec "say hi"
# or run codex and type a prompt; a normal reply means the connection works.

4 · Common pitfalls

• wire_api must be "responses": Codex 0.153+ no longer supports "chat" and only speaks /v1/responses. The international site https://quantoport.com serves /v1/responses; the CN site currently serves only /v1/chat/completions, so use the international site for Codex.
• env_key is just an environment-variable name; you can name it anything.
• Custom providers cannot reuse the reserved ids: openai, ollama, lmstudio.
• A project-level .codex/config.toml loads only when that directory is trusted.