🚀 LLM API & SLM + Microcontrollers = Fun Side Project! 🎤🤖

🚀 Exploring AI and Microcontrollers: My New Micropython Voice Assistant 🎤🤖

(For privacy protection, I have used a voice changer for the video audio.)

I’ve always loved AI and microcontrollers, though I am still very much a passionate beginner. A few years ago, I bought an M5Stack GO, but unfortunately, it ended up just sitting on the shelf collecting dust—I never really learned how to use it. Recently, I decided to give it another go and built a Micropython voice assistant for the M5Stack CoreS3 / ESP32-S3-Box!

This little device records speech, transcribes it using OpenAI Whisper, and then interacts via GPT—all by leveraging existing LLM APIs directly on this tiny hardware. The best part? No additional APIs or relay services needed!

💡 Features (so far!):

• ✔ Wi-Fi-enabled Voice-to-text transcription using OpenAI Whisper
• ✔ Chat capability with GPT & other LLM APIs (supports English, Japanese, and Mandarin)
• ✔ Not limited exclusively to OpenAI APIs—flexible and versatile!
• ✔ Fully Unicode-friendly (no troublesome urequest or API json parsing issues!)

Youtube video

⚠️ Note:

This project is currently just a prototype, built in a single day! I’d greatly appreciate feedback, advice, or even teammates who are also interested in embedded AI.

🔗 Let’s connect & build cool things together! 🛠️

I plan to open-source an improved version of the code soon, possibly next week.

Stay tuned!

#MicroPython #M5Stack #ESP32S3