Repolode
Submit
netease-youdao

netease-youdao/EmotiVoice

Open-source text-to-speech engine supporting English and Chinese with over 2000 voices and prompt-controlled emotional speech synthesis.

Open-source alternative to:ElevenLabs

View on GitHub
Stars
8,540
Forks
761
Open issues
137
Last commit
2 months ago

What is it?

EmotiVoice is a text-to-speech engine that generates English and Chinese speech with more than 2000 voices. Its main feature is emotional synthesis: speech can be rendered with emotions such as happy, excited, sad, and angry, controlled through a prompt.

The project provides a web interface for interactive use and a scripting interface for batch generation. It can also be run as a Docker image on a machine with an NVIDIA GPU, and includes an OpenAI-compatible TTS HTTP API. Inference takes phoneme-annotated text along with a speaker ID and style prompt, and the repository includes recipes for voice cloning using personal data, DataBaker and LJSpeech.

The project is built with Python and PyTorch and is released under the Apache-2.0 license. Support for more languages such as Japanese and Korean is listed as planned work.

Key features

  • 01Synthesizes English and Chinese speech with over 2000 voices.
  • 02Controls emotional style such as happy, sad, and angry through prompts.
  • 03Offers a web interface and a scripting interface for batch generation.
  • 04Provides an OpenAI-compatible TTS HTTP API.
  • 05Includes recipes for voice cloning with personal data.

Use cases

  • Generating emotional voiceovers for videos, audiobooks or games.
  • Adding a text-to-speech backend to an app through the OpenAI-compatible API.
  • Batch-producing Chinese or English audio from scripted text.
  • Researching prompt-controlled speech synthesis.

Quick start

  1. 1

    Run the EmotiVoice Docker image (needs an NVIDIA GPU)

    docker run -dp 127.0.0.1:8501:8501 syq163/emoti-voice:latest
  2. 2

    Open http://localhost:8501 in your browser to use the web interface

Things to know

  • The quickest Docker path requires an NVIDIA GPU and NVIDIA container toolkit setup.
  • The full installation needs downloading large pretrained model files, which can be troublesome.
  • The interactive page is provided under a separate User Agreement in addition to the Apache-2.0 license.

Topics

#text-to-speech#speech-synthesis#deep-learning#pytorch#emotion#voice-cloning#python#ai#emotivoice#multi-speaker

FAQ

What is EmotiVoice?
Open-source text-to-speech engine supporting English and Chinese with over 2000 voices and prompt-controlled emotional speech synthesis. EmotiVoice is a text-to-speech engine that generates English and Chinese speech with more than 2000 voices. Its main feature is emotional synthesis: speech can be rendered with emotions such as happy, excited, sad, and angry, controlled through a prompt.
Is EmotiVoice free to use? What license does it have?
EmotiVoice is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is EmotiVoice still maintained?
Updated within 6 months. The last commit to EmotiVoice was on Sep 3, 2026.
What can I use EmotiVoice for?
Generating emotional voiceovers for videos, audiobooks or games. Adding a text-to-speech backend to an app through the OpenAI-compatible API. Batch-producing Chinese or English audio from scripted text. Researching prompt-controlled speech synthesis.
How do I install EmotiVoice?
Run the EmotiVoice Docker image (needs an NVIDIA GPU): docker run -dp 127.0.0.1:8501:8501 syq163/emoti-voice:latest Open http://localhost:8501 in your browser to use the web interface
Where can I get EmotiVoice?
The source code and installation instructions are on GitHub: https://github.com/netease-youdao/EmotiVoice

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects