---
title: "Muse Code glossary: MSP, muse serve, ADE, approvals"
url: https://helicon.sh/glossary
description: "Plain definitions for the terms around Muse Code and agent tooling: the Muse Code Session Protocol, muse serve, ACP, ADEs, approvals, worktrees, subagents and more."
site: Helicon
---

# Glossary

Plain definitions for the vocabulary around Muse Code: the muse CLI, muse serve, the Muse Code Session Protocol, the Agent Client Protocol, agentic development environments, agent approvals, git worktrees, subagents, reasoning effort, TUIs, llms.txt and Tauri.

Each definition is self contained, so it reads correctly quoted on its own.

## Pages

- [Muse Code](https://helicon.sh/glossary/muse-code): Muse Code is Meta's AI coding agent, used through the muse command line tool on macOS, Linux and Windows. A plain definition, what it does, and how clients connect to it.
- [the muse CLI](https://helicon.sh/glossary/muse-cli): The muse command line tool is how you run Muse Code. What it installs, what muse login does, and why a GUI still needs it present.
- [the Muse Code Session Protocol](https://helicon.sh/glossary/msp): MSP is the JSON-RPC protocol that applications use to drive a Muse Code session over stdio, instead of parsing terminal output. What it carries and why it matters.
- [muse serve](https://helicon.sh/glossary/muse-serve): muse serve starts a Muse Code protocol host that applications can drive over stdio. What it is for, and how a GUI uses one host per workspace.
- [the Agent Client Protocol](https://helicon.sh/glossary/acp): ACP is a general protocol that lets editors talk to any conforming coding agent. How it differs from a Muse Code specific protocol, and what gets lost in the bridge.
- [an ADE](https://helicon.sh/glossary/ade): An ADE is the workspace you supervise coding agents from, the way an IDE is the workspace you write code in. What belongs in one and what does not.
- [agent approvals](https://helicon.sh/glossary/agent-approvals): An approval is the moment an agent asks before acting. The three Muse Code modes, what each one does, and why bypassing them is the wrong optimisation.
- [a git worktree](https://helicon.sh/glossary/git-worktree): A git worktree is a second working directory for the same repository. It is the cheapest way to stop two coding agents from editing the same files.
- [a subagent](https://helicon.sh/glossary/subagent): A subagent is a separate agent a main agent spawns to do part of the work in parallel. Why it saves wall clock time and why it is unreadable in a terminal.
- [reasoning effort](https://helicon.sh/glossary/reasoning-effort): Reasoning effort controls how much thinking a model does before answering. What it changes, what it costs, and when to turn it up.
- [a TUI](https://helicon.sh/glossary/tui): A TUI is a text user interface drawn inside a terminal. Why agents ship one first, and where the format runs out.
- [llms.txt](https://helicon.sh/glossary/llms-txt): llms.txt is a plain text file at a site's root that gives language models a clean summary and an index of the pages worth reading. What goes in one.
- [Tauri](https://helicon.sh/glossary/tauri): Tauri builds desktop apps from web UI using the operating system's own webview and a Rust core. Why that matters for a small, signed, auto updating app.