Onju Voice is an ESP32-S3-based board, carefully designed to be a drop-in replacement to the PCB in a Google Nest Mini (2nd gen), or used in one’s own custom enclosure, to create a powerful, privacy-focused, and expressive home voice assistant.

Onju Voice (top)

You can create an open source version of a home assistant using either cloud-based or local language models, text-to-speech, and transcription. While you can run powerful edge applications on the dual-core ESP32 with Espressif’s audio SDKs, the intended application that I provide examples for is offloading most AI capabilities to a server on your own network, letting the ESP32 handle audio processing.

Onju Voice (bottom)

All functionality of the Nest Mini is replicated in hardware, including Wi-Fi, Bluetooth, capacitive touch pads, mute switch, and six RGB LEDs. Onju Voice has a debug ribbon cable that can extend out of the device, and a USB port so that anyone can program it using the Arduino IDE or ESP-IDF.

Onju Voice in a Google Nest Mini (2nd gen)

Support & Documentation

• Schematics and read-only preview of ECAD
• Firmware, server application, and demos