Wally: The Portable Bilingual Voice Assistant

Overview
Wally is a warm and reliable voice assistant designed in the charming shape of a penguin, powered by the ESP32 platform. It serves as a personal AI companion that helps with reminders, enhances productivity, engages in daily conversations, and offers emotional support, all while being compact and portable.

Core Features

• Voice Wake-Up: Easily activate Wally with the wake word "Hi, Wally" to initiate interaction.
• Bilingual Voice Chat: Engage in fluent conversations in both English and Mandarin.
• Real-Time Streaming Responses: Experience instant replies from a sophisticated backend powered by XiaoZhi’s large language models (LLMs) such as Qwen and Deepseek.
• OLED/Display Feedback: Visualize replies on an OLED display with a variety of emojis for enhanced interaction.

Technical Overview

• Powered by: Integrates with the XiaoZhi cloud service alongside ESP32 firmware.
• Compatibility: Works seamlessly with DFRobot (UNIHiker) board.
• Backend Functionality: Utilizes XiaoZhi’s LLM framework for responsive interaction.
• Frontend: Comes with device firmware that can be flashed easily using XiaoZhi tools along with preset images.
• Wake Word Engine: Leverages ESP-SR for offline wake word recognition.
• Voice Synthesis: Features text-to-speech capabilities utilizing services like Volcano Engine or CosyVoice.
• Outer Shell: Creatively designed using Tynkercad and professionally 3D printed with the Bambu Lab P1.

Setup Instructions Summary

1. Power on Wally by flipping the switch on its back.
2. Access configuration mode by removing Wally's head and pressing designated buttons on the internal board.
3. Connect to Wally’s setup Wi-Fi network.
4. Input your Wi-Fi SSID and password via the configuration webpage.
5. Follow additional instructions to start using Wally once the network is configured.

Getting Started

• Say "Hi, Wally" to wake the assistant.
• Engage in conversation or request assistance.
• To power down, simply say any friendly goodbye.

Acknowledgements
Special thanks to Xiao Xia for their inspiring work on the XiaoZhi platform and the Xiao Zhi community for their support.

This project is a prototype and not yet available for general use or distribution, but updates will be shared soon. Feedback and contributions to enhance Wally are welcomed.