Skip to main content
This guide walks you through using Codex CLI locally and connecting it to the hboom API.

​
1. What is Codex CLI

Codex CLI is a local terminal AI coding assistant that helps you with:
  • Code generation
  • Code modification
  • Automated command execution
  • Project analysis
Once installed, use it directly in your terminal.

​
2. Prerequisites

​
1. Install Node.js

Recommended: Node.js 20 or higher Download: https://nodejs.org

​
2. Install Codex CLI

npm install -g @openai/codex
Verify installation: 
codex --version

​
3. Configure API

Codex uses a config file for API connection.

​
1. Config File Location

Default paths:
  • Windows:
C:\Users\<username>\.codex\config.toml
  • macOS / Linux:
~/.codex/config.toml

​
2. Basic Configuration

Edit config.toml: 
model = "gpt-4o"

[model_providers.default]
name = "default"
base_url = "https://betarouter.com/v1"
env_key = "OPENAI_API_KEY"
Fields:
  • model: model name to use
  • base_url: API address (must include /v1)
  • env_key: environment variable name for the API key

​
3. Set API Key

​
macOS / Linux

export OPENAI_API_KEY="your-api-key"

​
Windows (PowerShell)

setx OPENAI_API_KEY "your-api-key"

​
4. Start Codex

Navigate to your project: 
cd your-project
Launch: 
codex

​
5. Basic Usage

​
1. Interactive Mode

codex
Example prompt: 
Write a Python web scraper

​
2. Single Command

codex "Write a login API"

​
3. Specify a Model

codex -m gpt-4o "Optimize this code"

​
4. Image Input (Multimodal)

codex --image "./demo.png"
Analyze and process image content.

​
6. Troubleshooting

​
401 / 403 Error

Check:
  • Is the API Key correct?
  • Did the environment variable take effect?

​
404 Error

Check:
  • Is the API address correct?
  • Does it include the /v1 path?

​
Model Unavailable

Check:
  • Is the model name correct?
  • Does your account have access to this model?

​
7. Advanced Configuration

​
Multiple Models

model = "gpt-4o"

[model_providers.main]
base_url = "https://betarouter.com/v1"
env_key = "OPENAI_API_KEY"

[models]
gpt-4o = { provider = "main" }
gpt-4.1 = { provider = "main" }

​
8. Best Practices

  • Use environment variables for API keys (more secure)
  • Use consistent model naming for easier switching
  • Use separate configs per project
  • Regularly check usage and logs

​
9. Summary

After completing the setup, you can:
  • Use AI coding in your local terminal
  • Connect to a unified API service
  • Flexibly switch between models