explaingit

yeachan-heo/oh-my-codex

Analysis updated 2026-05-18

27,711TypeScriptAudience · developerComplexity · 3/5Setup · moderate

TLDR

A workflow layer for OpenAI's Codex CLI that breaks down complex coding tasks into structured steps: clarify intent, review plans, execute with single or parallel agents, and persist context between sessions.

Mindmap

mindmap
  root((repo))
    What it does
      Workflow structure
      Context persistence
      Multi-agent execution
      Terminal HUD display
    Key skills
      Deep interview
      Plan review
      Single-threaded loop
      Parallel team agents
    Use cases
      Architectural changes
      Multi-step features
      Complex refactoring
    Tech stack
      TypeScript
      Node.js
      OpenAI Codex
    Audience
      Codex CLI users
      Developers
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Break down architectural changes into structured steps with AI review and approval before execution.

USE CASE 2

Execute multi-step feature development with persistent memory across multiple Codex sessions.

USE CASE 3

Coordinate parallel AI agents to work on different parts of a complex task simultaneously.

USE CASE 4

Maintain project context and decision logs automatically in a .omx folder for reproducible workflows.

What is it built with?

TypeScriptNode.jsOpenAI Codex

How does it compare?

yeachan-heo/oh-my-codexbudibase/budibaselabring/fastgpt
Stars27,71127,89027,916
LanguageTypeScriptTypeScriptTypeScript
Setup difficultymoderatehardmoderate
Complexity3/53/53/5
Audiencedeveloperdeveloperdeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Requires OpenAI API key and Codex CLI installation, Node.js/TypeScript build step needed.

License could not be detected automatically. Check the repository's LICENSE file before use.

In plain English

OMX (oh-my-codex) is a workflow layer that sits on top of the OpenAI Codex CLI, which is a command-line coding agent. The README is clear that OMX does not replace Codex, Codex still does the actual agent work, but adds reusable roles, a standard workflow, hooks, runtime help, and a place to store plans and state, so day-to-day use of Codex feels more structured. The pitch is "start Codex stronger, then let OMX add better prompts, workflows, and runtime help when the work grows." You install it as a global npm package alongside Codex itself, then launch with a recommended command. From inside Codex you call canonical "skills" by typing keywords like $deep-interview to clarify a task, $ralplan to plan and review tradeoffs, $ralph to carry an approved plan to completion, $team for coordinated parallel execution, and $ultragoal for durable multi-goal launches. Project guidance lives in scoped AGENTS.md files, and OMX keeps plans, logs, memory, and runtime mode under a .omx/ directory. On macOS and Linux it can run the leader in a detached tmux session so it can manage HUD and runtime panes, launch behaviour is controlled with flags like --direct or --tmux, or an OMX_LAUNCH_POLICY environment variable. There is also an official Codex plugin layout in the repo with marketplace metadata, separate from the legacy install path. You would reach for OMX if you already use Codex and want a more opinionated runtime around it, repeatable workflows, specialist roles, and durable state across sessions. The README cautions that the recommended default is macOS or Linux with the Codex CLI, native Windows and the Codex App are not the default and may behave inconsistently. Requirements include Node.js 20+ and Codex login.

Copy-paste prompts

Prompt 1
Set up Oh My Codex in my project and configure it to use deep-interview mode for planning a major refactor.
Prompt 2
Show me how to use the $team skill to run parallel agents on different components of my codebase.
Prompt 3
How do I persist context between Codex sessions using OMX's .omx folder and $ultragoal for long-running work?
Prompt 4
Walk me through the $ralplan workflow: how do I review and approve an AI-generated plan before execution?
Prompt 5
Configure OMX to display a terminal HUD that shows the current step, logs, and memory state during execution.

Frequently asked questions

What is oh-my-codex?

A workflow layer for OpenAI's Codex CLI that breaks down complex coding tasks into structured steps: clarify intent, review plans, execute with single or parallel agents, and persist context between sessions.

What language is oh-my-codex written in?

Mainly TypeScript. The stack also includes TypeScript, Node.js, OpenAI Codex.

What license does oh-my-codex use?

License could not be detected automatically. Check the repository's LICENSE file before use.

How hard is oh-my-codex to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is oh-my-codex for?

Mainly developer.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Scan in gitsafehub Deploy in gitdeployhub yeachan-heo on gitmyhub

Verify against the repo before relying on details.