In this video, I build a fully working AI desk buddy using the ESP32-S3 and the Xiaozhi firmware.
Just an ESP32, an I2S microphone, an I2S amplifier, and a 128x64 OLED display.
What makes this build special:
• Custom animated face (blinking, idle movement, listening mode)
• Dynamic speaking mouth synced to audio
• Wake word detection
• I2S microphone + MAX98357A audio output
• OTA firmware support
• Runs fully on ESP32 hardware
I removed the default emoji UI and rewrote the display system to create a proper animated robot face with:
Idle personality movement
Listening side-eye animation
Random blinking
Dynamic mouth animation while speaking
This project is completely buildable by students and makers using affordable parts.
Hardware used:
• Seeed Xiao ESP32-S3
• MAX98357A I2S amplifier
• I2S MEMS microphone
• 128x64 OLED display
• 4Ω 3W speaker
The video ends with a working breadboard version, so you can replicate it exactly.
If you enjoyed this build, consider subscribing. More serious ESP32 projects coming soon.
Buy Xiao ESP32:
https://www.seeedstudio.com/XIAO-ESP3...
Code:
https://github.com/TechTalkies/Xiaozh...