oh-my-openagent

Good for

• Turn a request into finished work with one omo command: research across many sources, decks, backend and frontend code
• Add ultrawork (or ulw) to a prompt on hard tasks so it reads the codebase, plans and proves each step before acting
• Type mass ulw to turn the whole job into a parallel multi-agent graph, each part on a fitting model
• Keep memory in a git repo of markdown files so the agent remembers what you did before

Getting started

1. Run curl -fsSL https://get.omo.dev/install.sh | bash; on Windows PowerShell use irm https://get.omo.dev/install.ps1 | iex
2. Or install with a package manager: bun add -g omo-ai or npm i -g omo-ai
3. Open your project, run omo and describe the job
4. Use /thinking to set reasoning effort, and /resume or omo -c to continue earlier work

Before you install

The install script puts the native omo binary in ~/.local/bin. The npm package is omo-ai; the unrelated omo package is someone else's. To migrate from the OpenCode edition, run omo setup once.

Summarized by AI from the project docs; check the official docs before relying on it.