Py Xiaozhi, developed by Huangjunsen0406, is a Python-based AI & Memory server for the mcp platform. It connects open-source AI assistant ecosystems with physical hardware, acting as a bridge between Large Language Models and IoT devices. This free program supports Python versions 3.10 through 3.12, enabling voice AI interaction, multi-modal vision, and cross-platform deployment. It integrates with smart home platforms like Home Assistant, offering control over various devices and services.

What does Py Xiaozhi connect to?

Py Xiaozhi integrates with open-source AI assistant ecosystems through Model Context Protocol (MCP) integrations, functioning as an AI & Memory server. It connects Large Language Models such as Qwen, DeepSeek, and Doubao with physical hardware. The program also links with IoT devices and smart home platforms like Home Assistant. It supports Automatic Speech Recognition and Text-to-Speech services, using WebSocket, MQTT, UDP, and stdio as transport protocols for diverse client compatibility.

Core features and tools

Py Xiaozhi exposes several callable tools including music player, camera, screenshot, app management, weather, volume control, and Bazi fortune-telling. Through its IoT device integration, it also provides control over lights, temperature sensors, and countdown timers. These capabilities allow for interaction with both digital services and physical smart home environments, extending AI assistant functions to real-world applications.

Who is it for?

Py Xiaozhi targets developers, users seeking a complete XiaoZhi AI experience without specialized hardware, and individuals involved in AI hardware development and smart home scenarios. It offers a platform for experimenting with AI and IoT integrations across various operating systems and embedded platforms. This program suits those looking to combine advanced AI models with device control. Skip this if you are not interested in AI hardware development or smart home scenarios.

How to get started

The server authenticates using WSS/TLS encryption with certificate verification, dual protocol activation, and device fingerprint recognition. Device activation involves connecting to a Wi-Fi hotspot, accessing a configuration page, and entering a 6-digit device verification code on the XiaoZhi AI Console.

Use cases

• Real-time voice AI interaction and IoT device control, enabling natural language conversations and smart home device management.
• Multi-modal vision for image understanding and scene perception through camera capture and vision-language model integration.
• Cross-platform AI assistant deployment across Windows, macOS, Linux desktops, and ARM embedded platforms like Raspberry Pi.

Py Xiaozhi offers a versatile bridge between advanced AI models and the physical world, supporting diverse hardware and software integrations. Its ability to combine voice AI, vision, and IoT device control makes it a valuable tool for smart environments. The program supports multiple transport protocols and AI clients.