Codex CLI

Codex CLI is OpenAI's local terminal coding agent. It supports custom model providers through OpenAI-compatible endpoints. After connecting Codex CLI to WisGate, you can use WisGate models for code generation, project analysis, file edits, and debugging tasks in the terminal.

1. Install Codex CLI

Follow the official Codex CLI installation instructions first. After installation, verify that the command is available:

codex --version

2. Configure WisGate

Codex CLI usually reads ~/.codex/config.toml. Create the file if it does not exist.

model_provider = "wisgate"
model = "gpt-5.5"
model_reasoning_effort = "high"

[model_providers.wisgate]
name = "WisGate"
base_url = "https://api.wisgate.cn/v1"
env_key = "WISGATE_API_KEY"

Then set your WisGate API key in the shell:

export WISGATE_API_KEY="your_wisgate_api_key"

For persistent configuration, add the export line to ~/.zshrc, ~/.bashrc, or your shell profile.

3. Start Using

Navigate to your project directory and start Codex CLI:

cd /path/to/your/project
codex

Codex CLI will call the configured model through https://api.wisgate.cn/v1.

Use Cases

• Terminal development: Analyze projects, generate code, and modify files from the command line.
• Code review: Ask Codex CLI to inspect changes, summarize risks, and suggest fixes.
• Refactoring tasks: Split complex modules, adjust structure, and run verification commands.
• Team configuration: Standardize WisGate access with a shared config.toml template.

FAQ

Authentication fails

Confirm that WISGATE_API_KEY is set and that env_key matches the environment variable name.

The model is not available

Confirm that model is a model ID supported by WisGate. The OpenAI-compatible endpoint must be https://api.wisgate.cn/v1.