oh-my-openagent is now the omo binary: install paths, migration from OpenCode, and the naming traps
omo/lazycodex: The coding agent for tokenmaxxers;the one and only agent harness for complex codebases. For your Codex, for your OpenCode
At a glance
- What is it?
- code-yeongyu/oh-my-openagent, shipped as OmO, moved from an OpenCode plugin to a standalone omo command running on senpi, a fork of pi. This guide covers the install script and package routes, how updates and removal differ between them, the omo setup migration, and what ultrawork, Kibitzer memory and omo.jsonc actually do.
- Who is it for?
- OmO fits developers already living in OpenCode or LazyCodex who want multi-model orchestration, persistent memory and parallel agents behind one command, and who are comfortable with a tool that releases almost daily. It is a poor fit for teams that need a clear license answer up front, because the repository reports NOASSERTION and LICENSE.md has to be read directly.
- Can I use it commercially?
- Check first. The repository uses a licence we do not classify automatically, so read its LICENSE file before any commercial use.
- Is it still maintained?
- Yes. The repository received new commits within the last day.
- What is it written in?
- Mainly TypeScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.
Star history
- code-yeongyu/oh-my-openagent70,919 ★ · +2,887
| Repository | 2026-08-18 | 2026-11-15 |
|---|---|---|
| code-yeongyu/oh-my-openagent | 68k | 70.9k |
2,887 Stars in the last 89 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
The install script drops a native omo binary into ~/.local/bin
The primary route is a shell script followed by the first run:
curl -fsSL https://get.omo.dev/install.sh | bash
omoOn Windows the PowerShell equivalent is irm https://get.omo.dev/install.ps1 | iex. The script installs the native omo binary for your OS and CPU into ~/.local/bin, checks it against the release checksums, and falls back to GitHub Releases when the mirror is unreachable. Adding -s -- beta after bash picks the beta channel, and a value such as -s -- 5.1.1 pins a specific version. Windows CMD, PATH fixes, updating and uninstalling are covered at omo.dev/docs/install. One consequence: if ~/.local/bin is not already on your PATH, the omo line after the script will not be found until you fix that, which is why PATH gets its own page in the docs.
Three names for one tool, and an omo package that is not this one
The naming is the first trap. If you prefer a package manager, the install is bun add -g omo-ai or npm i -g omo-ai. The package is omo-ai. The plain omo package on npm belongs to someone else and is unrelated. Inside the repository, package.json is still named oh-my-opencode at version 5.1.59 and exposes five bin entries, oh-my-opencode, oh-my-openagent, omo-agent-toolkit, lazycodex and lazycodex-ai, all pointing at bin/oh-my-opencode.js. Its description still calls it a batteries-included OpenCode plugin. The project description lists omo and lazycodex together. For a reader this means search results, blog posts and old configs can refer to any of these names. Typing npm i -g omo by habit installs a different project entirely, so copy the package name exactly.
Update and uninstall commands depend on how you installed
The two install routes do not share an update path. A script install updates by running the install line again. A package-manager install updates with:
omo updateRemoval also splits. For the binary install it is rm ~/.local/bin/omo. For the package install it is bun remove -g omo-ai or npm uninstall -g omo-ai. If you install both ways on one machine, whichever omo comes first on your PATH wins, and updating the other copy changes nothing you can see. Pick one route per machine and write down which one, because the commands you need later follow from that choice and not from the version number.
omo setup carries an OpenCode or LazyCodex install across
Users coming from the OpenCode edition or LazyCodex run a one-time migration:
omo setupIt carries provider keys, custom providers, MCP servers, skills and model picks into the new install. Afterwards, omo doctor shows which task categories your providers can run and what the old install left behind, and omo doctor --models shows which model each category and agent runs and in what order. A model can be excluded everywhere with the exclude_models key. The full walkthrough sits in docs/guide/migrating-from-opencode.md. The doctor output matters because migration copies what it can but does not remove the old setup; leftovers stay until you clean them up, and a provider without the right subscription leaves some task categories unable to run.
ultrawork and mass ulw turn one prompt into planned or parallel work
Two prompt keywords change how OmO behaves. Adding ultrawork, or ulw, to a prompt makes the agent read the codebase before touching a line, plan, prove each step before moving on, check the result on the real surface, and stop when it is done. Typing mass ulw turns the job into a graph of agents, each part on the model that fits it, running in parallel and checked before anything is called done. In a session, /thinking sets how hard the model thinks, Shift+Tab cycles levels, /efforts sets reasoning effort on models with effort grades, and /resume or /sessions reopens past work. From the shell, omo -c continues the last session and omo -r opens the picker. The cost is tokens: running many agents across several models in parallel spends from every subscription you have signed in with /login, which covers Claude, ChatGPT, Kimi and GLM.
Kibitzer keeps memory as markdown files in a git repository
OmO's memory system, Kibitzer, stores memory in a git repository of markdown files. A small, cheap model sits beside the expensive main agent, reads what you have done before, and nudges the main agent before it repeats a mistake. Because memory is plain markdown under git, you can read, diff and edit what the agent remembers. Configuration follows a similar file-based approach: personal settings live in ~/.omo/omo.jsonc, a project can add its own .omo/omo.jsonc, and the closest file wins. Config, skills and prompts reload live. The trade is that a project-level omo.jsonc silently overrides your personal one inside that project, so a setting that works in one repository can appear ignored in another. Every key is listed in docs/reference/configuration.md.
A personal side project with near-daily releases and an unresolved license field
OmO describes itself as a personal side project kept going by sponsors, currently OpenGateway. The repository is not archived; the last push was on 2 November 2026, and releases v5.1.57, v5.1.58 and v5.1.59 landed on 1, 2 and 3 November 2026. The default branch is dev. GitHub reports the license as NOASSERTION. The repository holds LICENSE.md, CLA.md and THIRD-PARTY-NOTICES.md, while the Rust desktop crates in Cargo.toml declare MIT at the workspace level. For a company, that mix means reading LICENSE.md before adoption instead of assuming one license covers everything. Computer use, built on those senpi-desktop crates with backends for macOS, X11, Wayland, AT-SPI and Win32, is marked experimental. Bugs go through /bug inside a session, which shows you the exact report, including the recent engine log, environment and config, before sending.
Editorial conclusion
OmO fits developers already living in OpenCode or LazyCodex who want multi-model orchestration, persistent memory and parallel agents behind one command, and who are comfortable with a tool that releases almost daily. It is a poor fit for teams that need a clear license answer up front, because the repository reports NOASSERTION and LICENSE.md has to be read directly. Before installing, decide between the script install and the omo-ai package, since update and uninstall commands differ, and make sure you do not grab the unrelated omo package from npm.
Frequently asked questions
How do I install oh-my-openagent?
Run curl -fsSL https://get.omo.dev/install.sh | bash and then omo, or on Windows PowerShell use irm https://get.omo.dev/install.ps1 | iex. You can also install the omo-ai package with bun add -g omo-ai or npm i -g omo-ai.
How to update oh my openagent?
If you used the install script, run the install line again. If you installed the omo-ai package with bun or npm, run omo update.
How do I uninstall Oh My Open Agent?
For the script install, remove the binary with rm ~/.local/bin/omo. For a package install, run bun remove -g omo-ai or npm uninstall -g omo-ai.
What is oh my openagent?
oh-my-openagent, now called OmO, is a coding agent harness run as a single omo command. It runs on senpi, a fork of pi, and adds ultrawork planning, parallel agent graphs, multi-model sign-in and git-backed memory.
How to use oh my openagent?
Open your project, run omo and describe the job. Add ultrawork or ulw to a prompt for harder tasks, or mass ulw to split the job across a graph of agents.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/code-yeongyu-oh-my-openagent)Star history chart for your README
[](https://hysenlabs.com/projects/code-yeongyu-oh-my-openagent)Was this useful?