EmotiVoice is an open-source text-to-speech engine from NetEase Youdao, published on GitHub with around 8,500 stars. It focuses on expressive, controllable speech synthesis: beyond turning text into audio, it lets you steer the emotion and style of the voice with simple natural-language prompts.
EmotiVoice ships with more than 2,000 built-in voices covering a wide range of speakers, and it supports both English and Chinese synthesis. Its signature feature is prompt-based control — describe the delivery you want, such as happy, sad, angry, or excited, and the model adjusts its tone and speaking style accordingly. It also supports zero-shot voice cloning, where a short sample of reference audio is enough to synthesize new lines in that speaker’s voice. The repository includes the model code, training and inference scripts, and a web interface for trying voices quickly, which has made it popular for content creation, character voices, and interactive applications. The project remains actively maintained, with repository updates continuing into 2026.
EmotiVoice is free and open source under the Apache-2.0 license. You can download it from GitHub and run it on your own hardware with no subscription, usage fees, or cloud dependency.
EmotiVoice fits creators producing narrated video or audiobook content, game and animation teams needing expressive character voices, and developers who want a controllable, self-hosted TTS engine for Chinese and English applications.