Modern auditory rehabilitation faces significant cIRCAM:Galleryenges in speech discrimination within complex, noisy acoustic environments. The use of Augmented Reality interfaces based on "virtual sound objects" proposes the separation and selective enhancement of audio sources, while the Auracast standard (Bluetooth LE Audio) emerges as the ideal mechanism to distribute these independent streams with low latency. However, the advancement of such selective listening strategies is strictly limited by proprietary commercial ecosystems and a complete lack of open-source research platforms that adhere to the physical and power constraints of wearable devices. To bridge this gap, this work develops an open-source Auracast application on the Tiresias open-hardware platform, establishing an accessible "front-end" infrastructure for auditory interaction. The architecture was implemented on the Nordic nRF5340 SoC utilizing the Zephyr RTOS. Preliminary evaluations on a development kit successfully validated the protocol stack integration, demonstrating stream stability. Ongoing work focuses on porting the firmware to the Tiresias board and integrating the ADAU1787 audio codec, aiming to empirically quantify the end-to-end latency and energy efficiency of the embedded system.
