Get started · My Maestro AI

Install in one command

On a Mac or a Linux machine (on Windows, inside WSL), in a terminal:

curl -fsSL https://mymaestroai.ai/install.sh | bash

Everything goes into your home folder, without sudo: its own Node.js (checksum verified), the code from GitHub, and a service that starts by itself. Then the Welcome page opens to create your account, turn on two-factor sign-in and connect your first AI.

Prefer to read it first? curl -fsSLO https://mymaestroai.ai/install.sh, open install.sh, then run bash install.sh. The same script is install/install.sh in the repository.

Update your Maestro

5.5.1 is a security release (October 21, 2026). Node.js fixed high-severity security issues. A Maestro that was already running may still use a Node.js without those fixes, even after it updates itself, so updating takes two steps:

Press Update

On a computer: the ⋯ menu, then Update from GitHub. On the phone: the ⋮ menu, then Update hub. Keep your current channel: either way your Maestro ends up on 5.5.1 or newer. It restarts by itself.

Check its Node.js

Open Settings → Status → Security check. If it flags Node.js, follow the fix it shows there, which is written for your install, or ask your Maestro to do it (it asks before running anything). You are done when the Node.js line is no longer flagged.

Installed during the private preview? Start with the move to the public repository below. More: Security releases · Node.js security releases, October 2026

Four steps

From a fresh machine to a working team.

Install

Run the command above. It installs the service and opens the Welcome page, where you create your account and connect an AI. It runs under systemd or launchd and restarts itself after updates.

Meet your lead agent

Your first agent, Maestro, carries a complete map of the software. It helps you connect providers, create agents and fix anything the setup checklist flags.

Build your team

Give each agent a name, a model and a description of its role. The hub writes its identity, and you refine it as you get to know each other.

Take it everywhere

Install the phone app, turn on notifications, and connect the services your agents need. A private network such as Tailscale reaches the hub from anywhere.

What you need

A computer that stays on, running Linux or macOS (Windows: inside WSL). The installer brings its own Node.js.
At least one AI provider: a Claude or ChatGPT subscription, an API key from Anthropic, OpenAI, Google or Perplexity, or a local model.
Optional: Docker, for agent sandboxes.
Optional: an NVIDIA GPU or Apple silicon, for models that run entirely on your machine.

Works with

Claude: subscription or APIOpenAI: API or ChatGPT via CodexGoogle GeminiPerplexityllama.cppNInferLM StudioOllamaMCP serversGoogle WorkspaceGitHubTelegramBrave · Serper · Tavily · SearXNG

License and contact

My Maestro AI is source-available under the PolyForm Noncommercial license: free for personal, research and self-hosted non-commercial use. Commercial deployment, including use on behalf of clients, is available under a separate commercial license.

My Maestro AI is developed by Jeff Lunglhofer. The code is public at github.com/mymaestroai/mymaestroai. For commercial licensing or integration questions:

jeff@lunglhofer.com

Build your team.