fullive-ai/anima

Anima is an open-source system that tries to give your smart home devices genuine intelligence rather than just remote control. Most connected devices, even the expensive ones, sit passively waiting for commands. Anima runs on your local network as a kind of operating system for hardware, connecting your devices to an AI that can read the environment, understand what you prefer, and make decisions on its own. The name comes from the Latin word for "soul." The system has four main parts working together. The Brain is a central decision layer powered by any OpenAI-compatible AI service, including local models. It reads device state, your past preferences, and what the environment looks like right now, then figures out what to do. Skills are packages of domain knowledge for specific device types: the humidifier skill understands seasonal humidity ranges and how the humidifier should interact with your air conditioner, the lights skill can follow natural daylight patterns over the course of a day. Memory is where Anima stores what it has learned about your habits over time, so decisions get better as the system runs. Adapters handle the actual communication with real hardware. Anima currently supports Xiaomi Mi Home and MIoT devices. You can log in once via QR code and it automatically discovers your devices on the local network without you manually entering IP addresses or digging for device tokens. Other hardware protocols are planned but not yet included. Built-in skills cover lights, humidifiers, air conditioners, air purifiers, and speakers, plus a coordinator skill for decisions that involve multiple devices at once. There is also a skill creator that can generate a new custom skill from a plain-language description of what you want a device to do. You can add your own skills in a dedicated folder. The system includes a web dashboard for monitoring and control, a REST API for building on top of it, and a command-line tool. It requires Python 3.11 or newer. The project is Apache-licensed and open to community contributions for adding new hardware adapters and skills.