Everything you need.
Seven sections covering install through deep architecture. OMEGA is designed so most users are productive within five minutes of first launch. These docs are for when you want to understand what’s happening underneath.
Quick Start
OMEGA installs as a single Mac app. About five minutes from download to first command: no terminal, no Docker, no Homebrew, no external databases. Postgres, Redis, the agent runtime, and a bundled local embedder all ship inside the bundle.
1. Download OMEGA
Public macOS downloads are coming soon. The release channel remains closed until the signed artifact and operational checks are complete.
2. Sign in + grant permissions
On first launch OMEGA boots its bundled Postgres and Redis on isolated sockets, runs migrations, and prompts you to sign in with your OMEGA account. Then macOS will ask for the permissions OMEGA needs to control your Mac:
- Accessibility: for cross-app keyboard, mouse, and menu control
- Screen Recording: for screenshot capture during computer-use workflows
- Full Disk Access: for searching, reading, and writing files outside the sandbox
You can revoke any permission later in System Settings → Privacy & Security; OMEGA degrades cleanly when a permission is missing.
3. Choose your AI runtime
Open Settings → Models and connect at least one runtime. OMEGA supports BYOK providers such as OpenAI, Anthropic, Google, xAI, OpenRouter, and any OpenAI-compatible endpoint; Codex Subscription and Claude Subscription where supported; plus local providers (MLX, Ollama, LM Studio) at zero cost. Keys and subscription connection metadata are encrypted at rest with a per-machine secret stored in the macOS Keychain; they never appear in logs, audit trails, or the OMEGA app.
4. Issue your first command
Type or speak. OMEGA turns the request into clear work for the right agents and tools, then streams results back. Try:
- Inbox triage: “Read my last 10 unread emails, categorize by intent, draft replies for the urgent ones.”
- File organization: “Find every PDF in Downloads from this year and sort them into folders by topic.”
- Automation generation: “Build me an automation that watches my Stripe webhook and posts new orders to Slack.”