Installation & Your First Launch

This is the fastest way to go from a downloaded file to a working OOMU workspace. It takes about two minutes, and almost all of it happens on its own.

By the end you will have OOMU installed, your private workspace created on your Mac, and your local intelligence engine ready to answer its first question.

---

What you need

• A Mac with Apple Silicon (M1 or later). OOMU is a native Apple Silicon app.
• The OOMU application file; the OOMU.app itself.

That is the whole list. You do not need an account, an internet connection, or an API key to begin. OOMU is a desktop application, not a website; everything it sets up lives on your machine.

---

Step 1; Install OOMU

1. Open the .dmg you downloaded.
2. Drag OOMU into your Applications folder.
3. Eject the disk image.

That's it. OOMU is a single, self-contained application. There is nothing else to install.

The first time you open it: macOS may ask you to confirm that you want to open an application downloaded from the internet. This is the standard macOS Gatekeeper prompt. Choose Open. OOMU is code-signed, so you will only ever see this once.

---

Step 2; Launch and let OOMU set itself up

Open OOMU from your Applications folder.

On its first launch, OOMU quietly prepares your private workspace. You don't need to do anything, but here is what is happening behind the scenes:

• It creates your private data folder. OOMU sets up a dedicated folder under your Mac's standard Application Support location. This is where your assistants, conversations, settings, and installed capabilities live.
• It creates your secure local database. Your history and configuration are stored in an encrypted database on your machine. The encryption key is generated for your device and stored in the macOS Keychain; it never travels anywhere.
• It establishes your workspace identity. OOMU generates a private identity for your installation so that your saved work and any signed artifacts are bound to this machine.

All of this completes in the background within the first few seconds.

Where is my data? Your workspace lives in your user Library, under an OOMU-specific folder (ai.eldris.oomu.gpd). You never have to open it by hand, but it is yours; local, on disk, and under your control. See Privacy & Sandboxing for the full picture.

---

Step 3; Prepare your local intelligence engine

OOMU can think entirely on your Mac, without sending anything to the cloud. This is powered by a local model that runs inside OOMU.

The first time you use OOMU's local engine, it gets itself ready:

1. Open the Models area of OOMU.
2. OOMU checks for a local model in your workspace's models folder.
3. When a model is present and validated, OOMU reports it as ready. From then on, your assistant can answer locally; privately and at no per-question cost.

To keep things fast, OOMU pre-warms your default local model in the background when the app launches, so it's already loaded by the time you ask your first question. You can choose which local model is pre-warmed in the Models settings.

Heads up; the very first load of a model is the slowest moment you'll see. Loading a local model into memory takes a little longer than every subsequent answer. Pre-warming does this ahead of time, but if you switch to a model that isn't warm yet, expect a brief pause on its first reply. Once it's warm, local answers are fast.

If you'd rather connect a cloud provider (such as Anthropic, OpenAI, Google Gemini, and others) instead of, or alongside; the local engine, you can add your own API key in the Models settings at any time. See Smart Model Routing to learn how OOMU decides between local and cloud.

---

Step 4; The guided setup

The first time you open OOMU, it walks you through a short, resumable Setup flow so you leave with something that actually works; not just a saved configuration. You'll:

1. Choose how OOMU thinks. Either prepare a local model on this Mac, or connect a cloud provider with your own key. OOMU only ever offers sign-in methods it has fully implemented; it won't present an option that leads nowhere.
2. Grant only the Mac permissions your choices need. Files, Accessibility, Screen Recording, Automation, Notifications, and Keychain are requested one at a time, and only when a capability you picked actually requires one.
3. Optionally connect a service or two. Apple apps, Google Workspace, Slack, and one completion channel (like Signal or Telegram) can be connected now or later.
4. Run one real sample task, end to end, so you can watch OOMU work before you rely on it.

You can skip ahead and come back; the flow remembers where you left off.

---

Step 5; Find your way around

OOMU's sidebar has six doors, plus Settings and the Ledger in the footer:

---

Step 6; Ask your first question

Open Chat. On a fresh install, the chat greets you and offers a few starter cards; concrete things OOMU can do, phrased in plain words ("Get a weekly brief," "Summarize a folder," "Draft replies to my email"). Tap one to watch OOMU work, or just type a question below.

Watch the status line. OOMU tells you what it is doing in plain language; for example *"...is typing..."* for a direct answer, or *"...is planning local actions..."* when your request needs to work with files on your Mac. This is OOMU's router deciding the smartest way to handle your request. You'll learn to read these cues quickly, and you can dig into how they work in Smart Model Routing.

---

You're set up

You now have:

• ✅ OOMU installed in Applications
• ✅ A private, encrypted workspace on your Mac
• ✅ A working model, chosen through guided setup
• ✅ Your first conversation underway

Where to next

• Set up a Project: give OOMU a folder, some knowledge, and a data policy to work within.
• Set up a custom assistant: give OOMU a role, a name, and a thinking depth that matches your work.
• Add a custom capability: extend what your assistant can do with a .oomu package.
• Understand the design: see how OOMU runs on your desktop.