Codex

Prepare API

Before configuration, make sure you have the following information:

API Key

A string of your created API Key.

Base URL

Address: https://api.jalapeno-cloud.ai/v1

Important: Must include the trailing /v1

Model Name

Copy the full model name from the Model Marketplace, e.g., DeepSeek-V4-Pro

Install Codex

Download and install through Codex official channels. Codex CLI

Linux / macOS / WSL2

curl -fsSL https://chatgpt.com/codex/install.sh | sh

Or install via npm (requires Node.js 22+):

npm install -g @openai/codex

Windows (PowerShell)

powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"

Connect Codex

Codex routes requests through a custom model provider in your user-level config file ~/.codex/config.toml.

Important: model_provider and model_providers only take effect in ~/.codex/config.toml. Codex ignores them in project-local .codex/config.toml.

If you previously signed in with ChatGPT, log out first to avoid auth conflicts:

codex logout

Edit Configuration File

Create or edit ~/.codex/config.toml:

model = "GLM-5.1"
model_provider = "jalapeno"

[model_providers.jalapeno]
name = "Jalapeno Cloud"
base_url = "https://api.jalapeno-cloud.ai/v1"
env_key = "JALAPENO_API_KEY"
wire_api = "responses"
requires_openai_auth = false

Field Description

Field	Description
model	Default model name (e.g., GLM-5.1 / DeepSeek-V4-Pro / MiniMax-M2.7)
model_provider	Provider ID; must match the [model_providers.<id>] block below
name	Display name for the provider
base_url	API endpoint, must include trailing /v1
env_key	Environment variable holding your API key
wire_api	API protocol; use responses for third-party endpoints
requires_openai_auth	Set to false for OpenAI-compatible custom APIs

Set API Key

Codex reads the key from the environment variable named in env_key. Export it before running Codex:

export JALAPENO_API_KEY="your-api-key"

Add the export line to your shell profile (e.g. ~/.bashrc or ~/.zshrc) to persist it across sessions.

Switch Model

Change the model field in ~/.codex/config.toml, for example:

model = "DeepSeek-V4-Pro"

Verify Connection

With JALAPENO_API_KEY exported, run:

codex exec -m GLM-5.1 --sandbox read-only "Reply with exactly one word: hello"

A successful run prints a session header containing provider: jalapeno, followed by the model reply pong.

Use Model

Start the interactive CLI in your project directory:

codex