Codex CLI

Use models provided through ApiGo in Codex CLI for AI coding.

Install Codex CLI

Install Codex CLI globally with npm:

# Version 0.57.0 is recommended
npm i -g @openai/codex@0.57.0

Configure ApiGo API

Important: clear OpenAI environment variables before use

Before configuring, clear the following OpenAI-related environment variables so they do not override your ApiGo settings:

• OPENAI_BASE_URL
• OPENAI_API_KEY

Choose an integration method

@apigo/vibe-coding

@apigo/vibe-coding corresponds to the interactive terminal tool Agent CLI Switch. This flow writes and switches Codex config for you directly.

Follow the flow shown in the screenshots:

1. Open Agent CLI Switch v0.1.0.
2. In the Main menu, choose Configure tools.

3. In Select tool, choose Codex.

4. In Codex edit config, enter: API URL: https://maas.apigo.ai/v1 API Key: your ApiGo key
5. After you see the config-saved prompt, choose to apply the configuration immediately.

After that, run codex directly. Before you use it, confirm the currently supported Token Plan model IDs in the Model Catalog, then choose the model you need.

Manual configuration (Current)

1. Edit Codex’s config file at .codex/config.toml and add:

[model_providers.apigo]
name = "Chat Completions API"
base_url = "https://maas.apigo.ai/v1"
env_key = "APIGO_API_KEY"
wire_api = "chat"
requires_openai_auth = false
request_max_retries = 4
stream_max_retries = 10
stream_idle_timeout_ms = 300000

[profiles.apigo]
model = "your-model-id"
model_provider = "apigo"

2. Set APIGO_API_KEY to your platform key, then replace your-model-id with a model ID that is still supported in the Model Catalog.

3. Launch Codex CLI with the configured profile:

codex --profile apigo