Enables natural language control of a humanoid robot through an integrated voice agent and large language model.
This Python-based MCP server transforms the Baby Brewie robot into an intuitive, voice-controlled system. It integrates a dedicated voice agent for wake-word activation and speech recognition, sending commands to a large language model (LLM) via the Together API. The server also provides the LLM with context from pre-defined robot action groups, allowing for intelligent execution of complex tasks. Responses are voiced using gTTS and cached for reduced latency, while robust ROS communication ensures seamless robot interaction.