Tencent Hunyuan open-sources AuK, a unified speech generation and editing model

Tencent Hunyuan has officially released AuK, an open-source foundation model for unified speech generation and editing. It combines natural-language instructions with reference audio in one interface, supporting zero-shot TTS, instruction-controlled generation, content editing, Whisper-conversion, de-accent, timbre/style/emotion editing, speed/pitch control, enhancement, denoising, multi-speaker and music separation.

Alongside AuK, the team released AuK-Flash, which uses 4-step inference and is about 4.5x faster under matched conditions. Code, weights, and a demo are now live. The paper has been published on Hugging Face and the project is hosted on GitHub for developers and researchers to try and provide feedback.