Oh My Pi

v15.2.2Coding Agentsstable

⌥ AI Coding agent for the terminal — hash-anchored edits, optimized tool harness, LSP, Python, browser, subagents, and more

ai-agentai-coding-agentanthropicbunclaude
Share:
5,954
Stars
0
Downloads
0
Weekly
0/5

What is Oh My Pi?

Oh My Pi is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to ⌥ ai coding agent for the terminal — hash-anchored edits, optimized tool harness, lsp, python, browser, subagents, and more

⌥ AI Coding agent for the terminal — hash-anchored edits, optimized tool harness, LSP, Python, browser, subagents, and more

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • ⌥ AI Coding agent for the terminal — hash-anchored edits, op

Use Cases

Terminal coding agent
LSP integration
Optimized tool harness
can1357

Maintainer

LicenseMIT
Languagetypescript
Versionv15.2.2
UpdatedMay 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

NPM

npx -y @oh-my-pi/pi-coding-agent

Manual Installation

npx -y @oh-my-pi/pi-coding-agent

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Oh My Pi

Oh My Pi (omp) is a terminal-based AI coding agent with 32 built-in tools covering file editing, shell execution, LSP code intelligence, DAP debugging, browser automation, web search, image generation, and memory checkpointing. It supports multiple LLM providers out of the box — including Anthropic Claude, OpenAI, Gemini, and local models — and reads existing configuration from Cursor, Cline, Windsurf, and Copilot rule directories. Developers use it as a powerful alternative to Claude Code or similar agents when they want a feature-rich terminal agent with native LSP integration, sub-agent dispatch, and persistent memory across sessions.

Prerequisites

  • macOS, Linux, or Windows (PowerShell for Windows install)
  • Bun runtime (recommended) or Node.js 18+ for npm installation
  • An API key for your preferred LLM provider — authenticate via /login for OAuth providers (Anthropic, OpenAI, Google) or set provider-specific env vars
  • Optional: EXA_API_KEY, BRAVE_API_KEY, or TAVILY_API_KEY for web search tools
1

Install omp

Choose the installation method for your platform. Bun is the recommended runtime for best performance.

# macOS / Linux (curl installer)
curl -fsSL https://omp.sh/install | sh

# Homebrew
brew install can1357/tap/omp

# Bun (recommended)
bun install -g @oh-my-pi/pi-coding-agent

# npm
npx -y @oh-my-pi/pi-coding-agent

# Windows (PowerShell)
irm https://omp.sh/install.ps1 | iex
2

Authenticate your LLM provider

For OAuth-based providers (Anthropic, OpenAI, Google), use the /login slash command inside omp. For API-key providers, export the key before launching.

# Inside omp, type:
/login

# Or export directly:
export ANTHROPIC_API_KEY=sk-ant-your-key-here
export EXA_API_KEY=your-exa-key        # for web_search tool
export BRAVE_API_KEY=your-brave-key    # alternative search
3

Launch the interactive TUI

Start omp in interactive mode from your project directory. It will detect and load rules from .claude/, .cline/, .windsurf/, and similar directories automatically.

omp
4

Run a single prompt non-interactively

Use the -p flag to pass a single prompt and return, useful for scripting or CI pipelines.

omp -p "List all TypeScript files in src/ that import from deprecated modules"
5

Select tools and model

Restrict available tools with --tools and override the model with --model. Use --smol for cheap sub-agent tasks and --slow for deep reasoning passes.

omp --tools read,edit,bash --model anthropic/claude-sonnet-4-6

# Or cycle models mid-session with Ctrl+P or /model
6

Enable shell completions

Install shell completions for a better CLI experience with tab-completion for omp flags and slash commands.

# zsh
eval "$(omp completions zsh)"

# bash
eval "$(omp completions bash)"

# fish
omp completions fish > ~/.config/fish/completions/omp.fish

Oh My Pi Examples

Client configuration

omp is a standalone CLI agent, not configured through a JSON MCP config file. To use it as an MCP server source inside another tool, run it in RPC mode:

{
  "mcpServers": {
    "oh-my-pi": {
      "command": "npx",
      "args": ["-y", "@oh-my-pi/pi-coding-agent", "--mode", "rpc"]
    }
  }
}

Prompts to try

omp excels at multi-step coding tasks that combine editing, running code, and browsing.

- "Find all usages of the old API and refactor them to the new interface, then run the tests"
- "Open the browser, navigate to the project's GitHub releases page, and tell me what changed in the last version"
- "Debug the failing unit test, set a breakpoint, and inspect the variable state at the point of failure"
- "Search the web for the latest best practices for React Server Components and update our ARCHITECTURE.md"

Troubleshooting Oh My Pi

omp command not found after installation

If installed with Bun, ensure ~/.bun/bin is on your PATH: add `export PATH="$HOME/.bun/bin:$PATH"` to your shell rc file and reload with `source ~/.zshrc` or `source ~/.bashrc`.

LSP tools return no results or errors

The lsp tool requires a language server to be installed and reachable. For TypeScript, install typescript-language-server globally: `npm i -g typescript-language-server typescript`. Run omp from the project root so it can locate tsconfig.json.

web_search tool fails

The web_search tool requires at least one search provider API key: EXA_API_KEY, BRAVE_API_KEY, TAVILY_API_KEY, JINA_API_KEY, or similar. Export the key for your preferred provider before launching omp.

Frequently Asked Questions about Oh My Pi

What is Oh My Pi?

Oh My Pi is a Model Context Protocol (MCP) server that ⌥ ai coding agent for the terminal — hash-anchored edits, optimized tool harness, lsp, python, browser, subagents, and more It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Oh My Pi?

Install via npm with the command: npx -y @oh-my-pi/pi-coding-agent. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).

Which AI clients work with Oh My Pi?

Oh My Pi works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Oh My Pi free to use?

Yes, Oh My Pi is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "oh-my-pi": { "command": "npx", "args": ["-y", "@oh-my-pi/pi-coding-agent"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Oh My Pi?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides