Codex CLI

Codex CLI is OpenAI's terminal coding agent. It runs locally and safely modifies your codebase with patch-based editing (apply_patch), sandbox policies and approval modes.

• Official site: https://chatgpt.com/codex
• Repository: https://github.com/openai/codex

Prerequisites

• Codex CLI installed (npm i -g @openai/codex)
• Base URL: https://api.tokenhub.belivestudios.com
• API Key: create one in "API Tokens" on the BeLive Studio Console
• Model name: e.g. gpt-5.6-sol, exactly matching a model ID in the console model list

Edit the Config File

Edit ~/.codex/config.toml and add BeLive Studio as a custom model provider:

model_provider = "belive-studios"
model = "gpt-5.6-sol"

[model_providers.belive-studios]
name = "BeLive Studio"
base_url = "https://api.tokenhub.belivestudios.com"
env_key = "BELIVE_STUDIOS_API_KEY"
wire_api = "chat"

Then set the environment variable (add to ~/.zshrc or system environment):

export BELIVE_STUDIOS_API_KEY="sk-your-belive-studios-key"

Start Using

cd /path/to/your/project
codex

Use /model to switch models. For first-time use, set the approval mode to on-request so Codex asks before writing files or running commands.

FAQ

unauthorized errors? Make sure BELIVE_STUDIOS_API_KEY is exported and env_key in config.toml matches the variable name.

404 on requests? Check that base_url ends with /v1 and wire_api = "chat" is set.