Skip to main content

Getting Started

This guide walks through launching Maekon for the first time and confirming the agent is talking to a server.

1. Launch Maekon​

After installation, launch from the Applications folder (macOS), Start menu (Windows), or maekon command (Linux). The app starts minimized to the system tray — click the tray icon to open the main window.

2. Connect to a server​

Maekon pairs with a server endpoint that holds your knowledge graph and orchestrates suggestions. Two options:

Option A — Public demo server (api.oneshim.thengd.com)​

Suitable for evaluation. Click Sign in with email in the welcome screen, complete OAuth, and you'll be connected.

Option B — Self-hosted server​

If you've run the server locally or deployed it yourself, set the endpoint:

# Set custom server URL (replace with your own):
maekon config set server.api_url "https://your-server.example.com"
maekon config set server.grpc_url "https://grpc.your-server.example.com"

Then sign in. Tokens are stored in the platform keychain (Keychain on macOS, Credential Manager on Windows, libsecret on Linux).

3. Grant desktop permissions​

Maekon needs platform-specific permissions to capture context. The app will prompt on first run:

  • macOS — Screen Recording, Accessibility, Input Monitoring (Settings → Privacy & Security)
  • Windows — UI Automation (granted automatically on first capture)
  • Linux (X11) — runs without prompts; Wayland uses XWayland fallback

You can revoke any of these at any time — Maekon will degrade features rather than fail.

4. Confirm the suggestion stream​

After login, Maekon opens a gRPC stream to the server. The tray icon turns green and the main window shows recent activity in the timeline. To stop the stream click the tray icon → Pause monitoring.

5. Verify the build​

$ maekon --version
maekon 0.0.1-rc.3 (build: 2026-05-05 | commit: 628908505)

If the build date or commit looks unexpected, you may have an outdated install — re-run the quick install script.

What's next​