xiaozhi-esp32
Summary
An MCP-based chatbot | 一个基于MCP的聊天机器人
📖 Highlights
XiaoZhi is an MCP-based AI chatbot for ESP32 devices, enabling voice interaction with large models like Qwen/DeepSeek and multi-terminal control via the MCP protocol. It solves the problem of integrating AI voice assistants into embedded hardware with offline wake-up and cloud connectivity.
- Wi-Fi / 4G connectivity with offline voice wake-up via ESP-SR
- Streaming ASR + LLM + TTS for real-time voice interaction
- Supports OLED/LCD displays with emoji and battery management
- Device-side MCP for controlling speakers, LEDs, servos, and GPIO
- Cloud-side MCP for smart home, PC control, and knowledge search
- Multi-language support (Chinese, English, Japanese)
- Compatible with ESP32-C3, S3, P4 and 70+ open-source hardware boards
🤖 AI Deep Analysis
A solid choice for developers looking to build a custom chatbot on ESP32 using MCP, but not ideal for beginners or those needing out-of-the-box voice assistant features.
✅ Pros
- High performance on ESP32 hardware
- MCP-based architecture for modularity
- Active community with 27k+ stars
- Lightweight and efficient for IoT devices
⚠️ Cons
- Limited to ESP32 platform
- Requires understanding of MCP protocol
- Documentation may be sparse for beginners
🎯 Use cases
- IoT chatbot for smart home devices
- Embedded voice assistant on ESP32
- Educational projects for MCP protocol learning
⚖️ Comparison
Compared to other ESP32 chatbot frameworks like ESP-ADF or ESP-Skainet, this project focuses on MCP integration, offering more flexibility for custom chatbot logic but with a steeper learning curve. It is more lightweight than full-fledged solutions like Alexa or Google Assistant SDKs.
