VOICEVOX DIFY Plugin is a TTS (Text-to-Speech) plugin for the Dify platform that integrates with VOICEVOX, a free Japanese voice synthesis engine. This plugin allows Dify applications to convert text into natural-sounding Japanese speech using various voice models. This plugin is compatible with Dify 1.x.
This plugin works with
- VOICEVOX
- AivisSpeech
Download voicevox.difypkg in the release section and install it selecting Local Package File in the Plugins section on your Dify console.
- Ensure you have Python 3.12+ installed
- Install the required dependencies:
pip install -r ./voicevox/requirements.txt
- Copy .env.example to .env and configure:
INSTALL_METHOD=remote REMOTE_INSTALL_HOST=locahost REMOTE_INSTALL_PORT=5003 REMOTE_INSTALL_KEY=your_debug_key
Remote debugging information is available here.

-
If you get
plugin verificationerror, setFORCE_VERIFYING_SIGNATURE=false. See more details. -
Install FFMpeg (required just for Dify v1.0.0-beta.1)
This process is only required with v1.0.0-beta.1 since ffmpeg will be included in the newer version of docker-plugin_daemon docker image.
docker exec -it docker-plugin_daemon-1 apt-get update
docker exec -it docker-plugin_daemon-1 apt-get install -y ffmpeg# Execute main.py on your local development environment.
cd voicevox
python main.pyVoicevox plugin will appear in debugging mode.
Download the version suitable for your operating system. (dify-plugin-darwin-arm64 for Apple silicon mac). Put it in the root directory of this repository. https://github.com/langgenius/dify-plugin-daemon/releases
chmod +x dify-plugin-darwin-arm64
./dify-plugin-darwin-arm64 plugin package ./voicevox
voicevox.difypkg will be generated.
This project was inspired by and references code from the following sources:
- dify-voicevox-tts by uezo
- Basic TTS mpdel feature
- Update voice list based on the Voicevox speaker list API
- Tool to generate voice file
- Submit the plugin to the Dify plugin market place