What problem does it solve? Users need a conversational way to adjust their assistant's voice behavior and reach the right settings screens without hunting through menus or editing configuration files by hand. ## Core Features & Use Cases - Voice Configuration Updates: Change the TTS provider (ElevenLabs, Fish Audio, Deepgram, xAI), set voice IDs, adjust the push-to-talk activation key, and set the conversation timeout, with changes persisted to config and broadcast to the desktop client. - System Settings Navigation: Open macOS System Settings panes directly (Microphone or Speech Recognition privacy) to guide users through granting permissions. - In-App Settings Navigation: Jump the Vellum settings panel to a specific tab such as Voice, Models & Services, or Permissions & Privacy. - Use Case: A user says "switch my voice to Fish Audio" and the assistant validates the provider, updates the config, and applies the change immediately. ## Quick Start Ask the assistant to change the TTS provider to Deepgram and set the conversation timeout to 30 seconds.