xiaozhi-esp32 is a voice-interactive chatbot application for ESP32 microcontrollers that integrates large language models through the MCP protocol.
The project addresses the challenge of bringing conversational AI to embedded devices with limited resources. It works by connecting ESP32 boards to cloud-based large language models like Qwen or DeepSeek, using the MCP protocol to enable multi-terminal control and device integration. The system handles the full voice interaction pipeline: offline wake-word detection via ESP-SR, audio streaming with Opus codec, automatic speech recognition, language model inference, and text-to-speech synthesis. It supports multiple networking options including Wi-Fi, Ethernet, USB RNDIS, and 4G connectivity, allowing flexible deployment scenarios.
Adoption suits developers building voice-controlled IoT devices, smart home interfaces, or embedded AI assistants on ESP32 platforms. The tool is particularly valuable for projects requiring offline wake-word detection and real-time voice interaction without constant cloud dependency. It supports a wide range of ESP32 variants and includes features like speaker recognition, camera vision input on capable boards, OLED/LCD display output with emoji support, and battery management. The project provides extensive localization with 39 interface languages and includes both device-side MCP for controlling physical components and cloud-side MCP for extending model capabilities to smart home systems and other services.
Development activity shows sustained engineering focus on platform stability and compatibility. The project maintains support across multiple ESP32 chip families and has recently migrated to ESP-IDF version 6.0 as the primary target, with legacy version 5.5 retained only for documented older boards. Cryptographic code has been updated to use PSA Crypto standards, and the audio pipeline, packet validation, and hardware concurrency handling have been hardened. The maintainers validate releases across a matrix of board variants and provide detailed migration documentation for SDK updates.
