omp
A terminal coding agent with language-server and debugger tools built in.can1357/oh-my-pi
- Stars
- 37.5k
- Forks
- 4.2k
- Licence
- MIT
- Language
- TypeScript
- Last update
- 4 days ago
- Type
- CLI
What it is
omp is an open-source coding agent you run from the command line to have an AI model read, search, and edit code in your project. The README says it adds IDE-style features such as language server (LSP) support, a debugger integration, and subagents to a fork of the Pi coding agent. It supports many model providers, and the README claims 60+ providers and 31 built-in tools.
- Ask the agent to rename a symbol and have imports and re-exports updated across files
- Have the agent attach a debugger to a crashing binary and inspect the stack frame
- Run persistent Python or Bun code cells where the agent loads and charts data
- Split a refactor across subagents that each work in an isolated worktree
Great for
- Developers who want an AI agent working inside a terminal on their own codebase
- People who use multiple model providers and want one tool across them
- Developers debugging native, Go, or Python programs who want the agent to drive a debugger
- Users who want the agent to split larger jobs across parallel subagents
Not ideal for
- Non-programmers, since it is a developer tool that works in a terminal and on code
- People who need a graphical editor; the README describes a terminal agent, not an IDE replacement
- Teams needing a stable, long-term-supported release, since the README shows a v18.x release and says pull request policy is on trial
How to get it
- 1
Install with a package manager
You'll need: bun 1.4.4 or newer (for the recommended install)
Install with Bun (recommended)
bun install -g @oh-my-pi/pi-coding-agentInstall with Homebrew
brew install can1357/tap/ompInstall on macOS or Linux via script
curl -fsSL https://omp.sh/install | sh - 2
Run it
- Install omp using one of the commands in the README
- Enable shell completions if you want them (optional): eval "$(omp completions zsh)" for zsh
- Change into a project directory and run omp to start the agent
- Configure a model provider as described in the project docs
- 3
Get the most out of it
- Set up LSP servers per the README's LSP config docs so the agent uses your language tools
- Use the Alt+A Agent Hub shortcut to watch subagents while they run
- Add stream rules that inject course-corrections when the model goes off-script
- Enable shell completions for bash, zsh, or fish so model and flag names autocomplete
Downloads
Strengths
- Broad feature set documented in the README, including LSP, DAP debugging, and subagents
- Works with many model providers rather than a single vendor
- MIT licensed and open source with a public changelog and CI badge
Trade-offs
- The README's benchmark claims come from the project's own blog and are not independently verified here
- Several features (lsp, dap, subagents, stream rules) add complexity that a newcomer must learn
- Install paths vary by platform, and Alpine/musl users need extra system libraries
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| oh-my-pi (this) | 37.5k | CLI | intermediate | 4 days ago | Developers who want an AI agent working inside a terminal on their own codebase |
| OpenHands | 92.5k | App | intermediate | 4 days ago | Developers who want coding agents running on a server while their laptop is off |
| jcode | 21.4k | CLI | intermediate | 4 days ago | Developers who want several AI coding agents working on one repo at the same time |
| CowAgent | 47.8k | App | intermediate | 4 days ago | People who want a self-hosted personal assistant reachable from their everyday chat apps |
| ECC | 295k | Dev tool | intermediate | 4 days ago | Developers who already use Claude Code and want a shared workflow for planning, testing and review |
| hermes-agent | 260k | AI model / agent | intermediate | 4 days ago | Developers who want a personal agent they can host on their own server or machine |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.