This is a step-by-step tutorial on building a blazing-fast, fully offline AI chatbot using a Raspberry Pi 5 and the LLM8850 accelerator card. I’ll walk you through everything from hardware to software, and show how I moved the heavy work — ASR, LLM, and TTS — off the CPU and onto dedicated AI hardware. By the end of this video, you’ll have an offline AI chatbot that feels as fast as cloud-based models, without using the internet at all.

Link channel: https://www.youtube.com/@PiSugarStudio/videos

👉 LLM8850 Intergration Guide: https://github.com/PiSugar/whisplay-a…
👉 Whisplay Driver repository: https://github.com/PiSugar/Whisplay
👉 Whisplay AI Chatbot repository: https://github.com/PiSugar/whisplay-a…
👉 Get LLM8850: https://shop.m5stack.com/products/ai-…
👉 Get Waveshare Dual M.2 Hat: https://www.waveshare.com/pcie-to-2-c…
👉 Get Whisplay HAT: https://www.pisugar.com/products/whis…
https://www.amazon.com/dp/B0FPG8S6K6 https://www.tindie.com/products/pisug…
👉 Learn more about PiSugar: https://pisugar.com
👉 Raspberry Pi Imager: https://www.raspberrypi.com/software/