Skip to content

xiaozhi-esp32

Build an open-source AI chatbot on ESP32 hardware

Visit official site

What xiaozhi-esp32 is

The project is described as an open-source AI chatbot built on ESP-IDF that combines a large language model with hardware for a personal AI companion. It handles multilingual speech recognition and dialogue, and voiceprint recognition identifies the voice features of different users.

What you can do with it

  • Connect over Wi-Fi or ML307 Cat.1 4G
  • Wake the device offline through ESP-SR
  • Run streaming voice dialogue over WebSocket or UDP
  • Recognise five languages with SenseVoice and identify voiceprints with 3D Speaker

Example uses

  • A student builds an AI assistant that supports classroom teaching
  • A developer makes a question-answering robot for a specific industry
  • A hobbyist sets up a personal assistant for a smart home

Getting started

  • Visit the project GitHub repository and download the source code
  • Set up the development environment and install the ESP-IDF plugin
  • Choose a hardware board such as the ESP32-S3 development board
  • Configure network connection and speech recognition model parameters

Pros & cons

✓ What we like

  • Streaming voice dialogue runs over WebSocket or UDP
  • Voiceprint recognition uses 3D Speaker to tell users apart

! What to watch out for

  • The build starts from the GitHub source and ESP-IDF setup
  • A hardware board such as ESP32-S3 has to be chosen

FAQ

Which languages are listed?

The feature list names Mandarin, Cantonese, English, Japanese and Korean, five in total.

How is voice dialogue delivered?

Streaming voice dialogue is described over WebSocket or UDP protocols.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review