Repolode
Submit
code-yeongyu

code-yeongyu/oh-my-openagent

A TypeScript coding agent CLI where a single 'ulw' or 'mass ulw' keyword in your prompt makes it plan, delegate to multiple models, and verify the work.

Open-source alternative to:Cursor

View on GitHubWebsite
Stars
70,215
Forks
5,773
Open issues
1,569
Last commit
14 hours ago

What is it?

OmO is a command-line agent (`omo`) for running real work end to end, from research and documents to backends, frontends and code. It is built on senpi, a fork of the pi coding agent, and installs as a native binary via a shell or PowerShell script, or as the `omo-ai` npm/Bun package.

The core idea is keyword-triggered orchestration. Adding `ultrawork` or `ulw` to a prompt makes the agent read the codebase, plan, verify each step and stop when done. `mass ulw` splits a large job into a graph of agents that run in parallel on different models and are checked before completion.

It also includes a git-backed markdown memory system with a small secondary model that reminds the main agent of past mistakes, subscription-based sign-in for several model providers, bundled skills loaded on demand, and migration tooling from the OpenCode edition.

Key features

  • 01Keyword triggers (`ulw`, `mass ulw`) switch the agent into planned, verified execution.
  • 02Parallel agent graphs that route subtasks to different models.
  • 03Git-backed markdown memory with a secondary model that flags repeated mistakes.
  • 04On-demand skills, including browser use, loaded per task.
  • 05Sign-in for Claude, ChatGPT, Kimi and GLM subscriptions via `/login`.
  • 06`omo setup` migrates provider keys, MCP servers and skills from the OpenCode edition.

Use cases

  • Running a large migration across a codebase with parallel agents.
  • Researching many sources and producing a finished document or deck.
  • Building backend and frontend features from a single prompt.
  • Combining expensive frontier models with cheaper ones to control token spend.

Quick start

  1. 1

    Install OmO via npm

    npm i -g omo-ai
  2. 2

    Open your project and run OmO

    omo
  3. 3

    Describe the job, adding ulw to your prompt for hard tasks

Things to know

  • Uses the SUL-1.0 license, not a standard open-source license, so check its terms before commercial use.
  • Running it requires signing in to model subscriptions or providing provider keys.
  • Computer use is experimental.

Topics

#ai-agents#coding-agent#llm#orchestration#cli#typescript#claude#openai#ai#anthropic

FAQ

What is oh-my-openagent?
A TypeScript coding agent CLI where a single 'ulw' or 'mass ulw' keyword in your prompt makes it plan, delegate to multiple models, and verify the work. OmO is a command-line agent (`omo`) for running real work end to end, from research and documents to backends, frontends and code. It is built on senpi, a fork of the pi coding agent, and installs as a native binary via a shell or PowerShell script, or as the `omo-ai` npm/Bun package.
Is oh-my-openagent free to use? What license does it have?
oh-my-openagent is open source under the Other license. Check the license terms before using it in commercial products.
Is oh-my-openagent still maintained?
Actively maintained. The last commit to oh-my-openagent was on Oct 20, 2026.
What can I use oh-my-openagent for?
Running a large migration across a codebase with parallel agents. Researching many sources and producing a finished document or deck. Building backend and frontend features from a single prompt. Combining expensive frontier models with cheaper ones to control token spend.
How do I install oh-my-openagent?
Install OmO via npm: npm i -g omo-ai Open your project and run OmO: omo Describe the job, adding ulw to your prompt for hard tasks
Where can I get oh-my-openagent?
The source code and installation instructions are on GitHub: https://github.com/code-yeongyu/oh-my-openagent

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects