Every room can hear you and every room can answer. The voice layer is fully local and trilingual; the music layer reaches every speaker; the screens are all first-class citizens of the same system.
Music Assistant is the jukebox: Spotify as the primary catalogue with Tidal enabled beside it as a standing fallback, playing to any room or combination. The voice path resolves fuzzy requests (“play chilled house”), supports “artist radio” that queues similar music, verifies that playback actually started before claiming success, and retries through the other provider if the first one misbehaves.
Films work the same way: “play The Rookie on the living room TV” launches Jellyfin’s Moonfin app on the Apple TV or Chromecast, waits until the play position is actually moving, and only then reports success — pause, resume and stop are verified against the same session. Replies come back in three voices — a Thai voice we cloned with F5-TTS-THAI, and Piper voices for English and Mandarin — chosen by a small router of our own that reads the script of the reply. The full story of the voice layer is its own page: a voice assistant that speaks the whole family’s languages.