Jieyue Xingchen Releases StepAudio 2.5 TTS Model, Making AI Voice Expression More Context-Aware

Jieyue Xingchen today released its new voice generation model StepAudio 2.5 TTS, which the official claims allows everyone to be a dubbing director. It is the first to integrate context understanding into the entire voice generation process. Through dual control of global context and inline context, combined with zero-shot voice cloning technology, AI no longer just reads text but can perform it.

The model has three core capabilities: global context control allows customizing the emotional tone, character state and scene atmosphere of the entire speech; inline context control can adjust details like tone, rhythm and pauses of a single sentence; zero-shot cloning retains the target voice while supporting flexible adjustment of emotion and style, enabling the same voice to present diverse expressions.