OpenAI Codex & API Integration

OpenAI Codex CLI - Quick Setup

OpenAI Codex is an AI coding agent included in ChatGPT Plus, Pro, Business, Edu, and Enterprise plans. Configure it to generate emoji-free code.

1. Install Codex CLI

npm i -g @openai/codex

2. Download nomoji rules

curl https://nomoji.dev/api/template/default/openai-codex -o ~/nomoji-codex.md

3. Configure Codex system instructions

codex config --system-instructions ~/nomoji-codex.md

Note: Check Codex documentation for the exact configuration command as it may vary

4. Verify it works

Test with a simple prompt:

codex "Create a hello world Express server"

Output should be emoji-free

Project-Specific Configuration

Add nomoji rules to your project:

# In your project root
mkdir -p .codex
curl https://nomoji.dev/api/template/default/openai-codex -o .codex/nomoji.md

# Configure per-project (if supported)
codex config --local --system-instructions .codex/nomoji.md

# Commit to version control
git add .codex/
git commit -m "Add Codex nomoji configuration"

How It Works

Once configured, Codex will:

Generate code without emojis in any context
Use plain ASCII for console output and progress indicators
Create professional documentation and README files
Write clean commit messages and code comments
Follow emoji-free guidelines across all file operations

OpenAI API - Integration Guide

Use nomoji rules as system instructions when calling the OpenAI API for code generation.

1. Download system instructions

curl https://nomoji.dev/api/template/default/openai -o nomoji-system-prompt.md

2. Load and use in your code

import OpenAI from 'openai';
import fs from 'fs';

const client = new OpenAI();
const systemPrompt = fs.readFileSync('nomoji-system-prompt.md', 'utf-8');

const response = await client.chat.completions.create({
  model: 'gpt-4',
  messages: [
    { role: 'system', content: systemPrompt },
    { role: 'user', content: 'Create a Node.js server' }
  ]
});

3. Or fetch dynamically

const response = await fetch(
  'https://nomoji.dev/api/template/default/openai'
);
const systemPrompt = await response.text();

// Use in OpenAI API calls
const completion = await client.chat.completions.create({
  model: 'gpt-4',
  messages: [
    { role: 'system', content: systemPrompt },
    { role: 'user', content: userPrompt }
  ]
});

Recommended Models

For best emoji-free results, use reasoning models which better follow constraints:

o1 - Best at following strict no-emoji instructions
o3-mini - Cost-effective with strong instruction following
o3-mini-high - Enhanced reasoning for code generation
gpt-4 - Works well but may occasionally include emojis
gpt-4-turbo - Fast with good instruction adherence

Note: GPT-4o was designed to use more emojis. For code generation, prefer the models listed above.

API Best Practices

Always include the system prompt in your API calls
Use temperature=0 for more deterministic, professional output
Consider caching the system prompt to reduce token usage
Monitor completions and provide feedback if emojis appear
Use reasoning models for stricter constraint adherence

Troubleshooting

Codex: Rules not being applied?

Verify Codex is installed: codex --version
Check your ChatGPT plan includes Codex access
Re-apply configuration with updated rules
Try setting rules per-project instead of globally

API: Emojis still appearing in responses?

Verify system prompt is included in messages array
Try switching to o1 or o3-mini models
Set temperature to 0 for more consistent output
Check your system prompt is the first message in the array
Refresh rules: curl https://nomoji.dev/api/template/default/openai