Skip to main content

AI Audio Dongle

FrontRetro
info

Depending on the configuration and following revisions, the board's features are subject to change. For detailed information on hardware specifications, please visit www.seco.com.

The Edge AI Audio Dongle is an advanced USB-C voice interaction peripheral designed to enable natural, reliable human-machine communication in demanding industrial and commercial environments.

Equipped with a six-element MEMS microphone array arranged in a 3+3 configuration, the device leverages beamforming, noise reduction, and voice isolation to create a focused audio interaction zone. Its directional listening capabilities help filter background noise and surrounding conversations, enabling accurate speech recognition and interruption-aware interactions with AI assistants, even in environments with multiple operators.

The device features an integrated mono speaker and an optional 5MP autofocus camera with illumination LED, supporting both voice-based and multimodal interactions. Audio processing is performed locally on the host system, allowing OEMs to introduce advanced voice capabilities without redesigning their existing HMI platforms.

Featuring a compact 175 × 30.5 × 21 mm design, USB-C connectivity, 5V power supply, and IP65 protection rating, the Edge AI Audio Dongle is designed for easy integration into industrial systems. Compatibility with Clea OS, based on Yocto Linux, further supports deployment across smart vending machines, industrial automation, interactive kiosks, and healthcare applications.

Technical Documents​

At the following Links you will find documentations like Manual and Rev. specific documents.

Manual​

[Manual] AI Audio Dongle (Coming soon)

Datasheet​

[Datasheet] AI Audio Dongle

Product page​

AI Audio Dongle

Software Resources​

Get started with AI Audio Dongle
This guide is designed to walk you through everything you need to know to get started: from building Clea OS (Yocto) with the bdSound S2C ALSA plugins, to runtime audio capture/playback, evaluation license limitations, and acoustic validation.