HubLensLLMXiaoMi/xiaomi-miloco
XiaoMi

xiaomi-miloco

View on GitHub
2,549

// summary

Xiaomi Miloco is an open-source smart home solution that utilizes on-device large language models to integrate and control IoT devices. By leveraging camera data streams, the system enables natural language interaction for complex home automation and event analysis. It prioritizes user privacy by performing visual understanding and task planning locally on the user's hardware.

// technical analysis

Xiaomi Miloco is an innovative smart home framework that leverages a self-developed on-device Large Language Model to transform home automation through natural language interaction. By integrating visual data from Xiaomi Home cameras with an LLM-based planning and understanding engine, it enables users to define complex, creative home rules without manual programming. The architecture prioritizes privacy and security by performing all visual processing and model inference locally, effectively bridging the gap between raw sensor data and intelligent, intent-driven device control.

// key highlights

01
Enables natural language interaction for setting complex home rules and controlling IoT devices, replacing traditional rigid configurations.
02
Utilizes the Xiaomi MiMo-VL-Miloco-7B on-device model to perform local video understanding, ensuring user privacy by keeping data off the cloud.
03
Splits home automation tasks into planning and visual understanding stages to provide accurate, context-aware responses to user queries.
04
Integrates directly with the Xiaomi Home ecosystem to allow seamless retrieval and execution of device commands and automated scenes.
05
Supports customized content delivery through Xiao Home notifications, enhancing the feedback loop between the AI and the user.

// use cases

01
Natural language control of Xiaomi Home IoT devices and scenes
02
On-device visual understanding and event analysis using camera streams
03
Customizable home automation rules defined through conversational AI

// getting started

To begin, ensure your system meets the hardware requirements, including an NVIDIA GPU with at least 8GB of VRAM and a Linux-based environment or WSL2 on Windows. You can initiate the setup by running the provided one-click installation script via the command line, which handles the necessary Docker configurations. Once installed, refer to the project's usage documentation to begin configuring your local smart home environment.