Spoken dialogue models have significantly advanced intelligent human-computer interaction, yet they lack a plug-and-play full-duplex prediction module for semantic endpoint detection, hindering seamless audio interactions. In this paper, we introduce Phoenix-VAD, an LLM-based model that enables streaming semantic endpoint detection. Specifically, Phoenix-VAD leverages the semantic comprehension capability of the LLM and a sliding window training strategy to achieve reliable semantic endpoint detection while supporting streaming inference. Experiments on both semantically complete and incomplete speech scenarios indicate that Phoenix-VAD achieves excellent and competitive performance. Furthermore, this design enables the full-duplex prediction module to be optimized independently of the dialogue model, providing more reliable and flexible support for next-generation human-computer interaction.
@article{arxiv.2509.20410,
title = {Phoenix-VAD: Streaming Semantic Endpoint Detection for Full-Duplex Speech Interaction},
author = {Weijie Wu and Wenhao Guan and Kaidi Wang and Peijie Chen and Zhuanling Zha and Junbo Li and Jun Fang and Lin Li and Qingyang Hong},
journal= {arXiv preprint arXiv:2509.20410},
year = {2025}
}