A user demonstrates a local AI assistant setup running the Qwen 3.6 35B model on LM Studio, projected to an ESP32 W-10 "Rune Board" via a local Xiaozhi server.
The system utilizes a modded flow connecting Mi agents and a Telegram bridge through a Tailscale tunnel to the hardware. Capabilities include continuous listening, file transfer, music generation, and personal assistant tasks like class tracking and life logging. The user highlights the Qwen model's competence compared to cloud alternatives and notes future plans for LoRa connectivity.