Getting Started

Table of Contents

  1. Quick Start
    1. 1. Launch Maestro
    2. 2. Select a Project
    3. 3. Configure Sessions
    4. 4. Launch
    5. 5. Work in Parallel
  2. Session Management
    1. Adding Sessions
    2. Closing Sessions
    3. Changing Mode
    4. Assigning Branches
  3. Git Worktree Isolation
    1. How Worktrees Work
  4. Quick Actions
  5. Status Indicators

Quick Start

1. Launch Maestro

Open the application after installing it.

2. Select a Project

Choose a project directory — ideally a git repository. Maestro uses git worktrees to isolate each session, so a git repo is required for full functionality.

3. Configure Sessions

In the sidebar, set up your sessions:

  • Number of terminals — Choose between 1 and 6 concurrent sessions
  • AI mode — Select the AI assistant for each session (Claude Code, Gemini CLI, OpenAI Codex, or Plain Terminal)
  • Branch assignment — Pick a branch for each session to work on

4. Launch

Click “Launch” to start all sessions. Each session opens in its own worktree with the AI ready to work.

5. Work in Parallel

Each session now runs independently. You can:

  • Watch AI agents work on different tasks simultaneously
  • Respond to input prompts as they appear
  • Monitor status indicators for each session
  • Use quick actions to trigger common commands

Session Management

Adding Sessions

Click the floating + button to add a new session to the grid. The grid layout adjusts automatically.

Closing Sessions

Click the x on the session header to close it. The associated worktree is cleaned up automatically.

Changing Mode

Use the mode dropdown in the session header to switch between AI assistants (or plain terminal) for any session.

Assigning Branches

Select a branch from the branch dropdown in each session’s configuration. When you assign a branch:

  1. Maestro creates a worktree at ~/.claude-maestro/worktrees/{repo-hash}/{branch}
  2. The session’s terminal opens in that worktree directory
  3. All file changes are isolated to that worktree
  4. The worktree is cleaned up when the session closes

Git Worktree Isolation

Worktree isolation is the core of Maestro’s parallel workflow. Here’s how it works:

Without Maestro: All AI sessions share one working directory. If Session A modifies app.tsx while Session B is also editing it, you get conflicts and broken code.

With Maestro: Each session gets a separate copy of the repository (a git worktree). Session A’s changes to app.tsx are completely invisible to Session B until you merge the branches.

How Worktrees Work

  • Each worktree is a full checkout of the repo at a specific branch
  • Worktrees share the same .git data (history, objects) but have independent working files
  • Changes in one worktree don’t affect others
  • When you merge a branch, the changes flow through git as normal

Quick Actions

Each session can have quick action buttons that send commands to the AI:

Action What It Does
Run App Tells the AI to run the application
Commit & Push Tells the AI to commit and push changes
Custom Configure your own prompts

Quick actions are configurable per session and can be customized in the sidebar.


Status Indicators

The MCP integration gives you real-time visibility into what each AI session is doing:

Status Meaning
Idle AI is ready and waiting for input
Working AI is actively processing a task
Needs Input AI is waiting for your response
Finished AI has completed its task
Error Something went wrong

Status updates appear as colored indicators in the session grid, so you can see at a glance which sessions need your attention.


Built with love by Jack. Licensed under MIT.

This site uses Just the Docs, a documentation theme for Jekyll.