XiaoZhi AI Chatbot Documentation Center | XiaoZhi.Dev

XiaoZhi AI Chatbot Documentation Center | XiaoZhi.Dev

📖 XiaoZhi AI Documentation Center

XiaoZhi AI is an open-source intelligent voice robot based on ESP32-S3 development, integrating wake word detection, AI conversation, device control, and multi-protocol communication capabilities. This documentation center provides complete technical guides from hardware assembly to AI integration.

Project Features:

  • 🎙️ Wake Word Detection: Supports 26 wake words including “Hello XiaoZhi”, <200ms response
  • 🧠 AI Integration: Supports multiple LLMs including DeepSeek, GPT, Ernie Bot
  • 🏠 IoT Control: Integrates MQTT, MCP protocols for smart home control
  • 🔧 Open Source Hardware: Based on ESP32-S3, complete open-source solution

🚀 Quick Start

Beginner’s Guide


📚 Documentation Navigation

🛠️ User Guide

Complete tutorials for users

💻 Development Documentation

In-depth technical guides for developers

🔧 ESP32 Development Guide

Complete ESP32-S3 platform development tutorial

🤖 AI Feature Capabilities

AI technology integration and capability overview

  • 🎯 Voice Processing: Local wake word + cloud recognition hybrid solution
  • 🧠 LLM Integration: Support for DeepSeek, GPT, Qwen and other models
  • Edge Inference: TensorFlow Lite lightweight model integration
  • 🎵 Voice Synthesis: Multi-engine TTS and emotional voice output

🛠️ Technical Features

Hardware Platform

  • Main Controller: ESP32-S3 dual-core 240MHz, 16MB Flash + 8MB PSRAM
  • Audio Processing: INMP441 digital microphone + MAX98357A digital amplifier
  • Display Output: SSD1306 OLED display + RGB status lights
  • Network Communication: Wi-Fi 2.4GHz + 4G Cat.1 communication (optional)

Software Architecture

  • Development Framework: ESP-IDF v5.3.2 + Arduino compatible
  • AI Engine: Espressif Wake Word Engine + cloud LLMs
  • Communication Protocols: WebSocket + MQTT + MCP Protocol
  • Audio Encoding: 16kHz PCM + Opus compression transmission

📊 Performance Metrics

Feature ModulePerformance MetricsNotes
Wake Word Detection<200ms latency, >99% accuracyLocal offline processing
Speech Recognition<1s latency, >95% accuracyChinese recognition accuracy
AI Conversation<3s response, supports 5+ LLMsDeepSeek recommended
Device Control<100ms command responseLocal + cloud hybrid
Power Management5mA standby, 150mA activeSmart power optimization

🗂️ Documentation Index

Documentation CategoryDocument NameMain ContentUpdate Time
User GuideHardware Setup GuideESP32-S3 assembly, wiring diagrams, component lists2025-03-19
User GuideFirmware DownloadPre-compiled firmware, flashing tools, configuration guide2025-03-18
User GuideNetwork ConfigurationWi-Fi setup, troubleshooting, advanced settings2025-03-18
User GuideFeature TutorialVoice interaction, device control, personalization settings2025-03-18
User GuideESP32 Budget VersionLow-cost ESP32 development board construction solution2025-03-18
User GuideFAQUsage issues, troubleshooting, technical support2025-03-18
Development DocumentationESP-IDF Environment SetupDevelopment environment configuration, compilation toolchain installation2025-03-06
Development DocumentationWebSocket ProtocolCommunication protocol specifications, message format definitions2025-03-06
Development DocumentationMCP Protocol SpecificationModel Context Protocol interaction flow2025-03-20
Development DocumentationMCP Usage GuideSpecific applications of IoT device control2025-03-20
Development DocumentationMQTT+UDP ProtocolControl channel and audio channel hybrid communication2025-03-20
Development DocumentationEmoji Emotion DisplayLLM emotion state expression protocol2025-03-06
ESP32 DevelopmentTechnical SpecificationsESP32-S3 hardware architecture, performance parameters2025-09-25
ESP32 DevelopmentProgramming GuideGPIO control to complex system development2025-09-25
ESP32 DevelopmentAdvanced Features4G communication, AI inference, multimodal interaction2025-09-25
ESP32 DevelopmentTroubleshootingIssue diagnosis, solutions, debugging techniques2025-09-25
AI FeaturesAI Feature IntegrationVoice processing, LLM integration, edge inference2025-09-25

🔗 Related Resources

Community Resources

Technical Support: