Download
wukong-robot is a Chinese voice assistant / smart speaker project built to let makers and hackers design highly customizable voice-controlled devices. It combines wake-word detection, automatic speech recognition, natural language understanding, and text-to-speech into a single framework aimed at the Chinese-speaking ecosystem. The project is positioned as a simple, flexible, and elegant platform that can run on devices like Raspberry Pi and other Linux-based boards, making it suitable for DIY smart speakers and home-automation hubs. It supports multi-turn conversational capabilities powered by ChatGPT or other large language models, letting users have continuous dialogues rather than one-shot commands. The project emphasizes extensibility: there is a plugin ecosystem (wukong-contrib) where developers can add new skills such as controlling smart-home devices, querying services, or performing custom actions.
Features
- Chinese-language smart speaker / voice assistant framework for makers and hackers
- Multi-turn ChatGPT-powered conversation support for more natural dialogues
- Plugin architecture via the wukong-contrib ecosystem to add custom skills and integrations
- Designed to run on Raspberry Pi and similar boards, with Docker images available for easier deployment
- Experimental brain–computer interface wake-up support as part of its research-oriented features
- Integration potential with ASR, TTS, and home-automation platforms (e.g., HomeAssistant and similar systems)
Project Samples
