Qwen-TTS backend for narration

A high-performance neural speech synthesis engine optimized for long-form narration and natural prosody.

The Qwen-TTS backend leverages the Qwen-Audio architecture to deliver studio-quality narration with minimal latency. It supports multi-speaker synthesis and fine-grained emotional control (including stress and intonation patterns) across 10+ languages. By utilizing a transformer-based decoder, the system maintains consistent vocal character over extended scripts, making it the primary choice for automated audiobook production and high-fidelity voiceovers.

https://github.com/QwenLM/Qwen-Audio