Notice

This is not working with Dify 1.0.x
I am developing Dify plugin version of the feature.

What is this script?

The script installs dify-voicevox-tts into Dify with the list of voice models supported by Voicevox or AivisSpeech API.

Usage

Preparations

• Open Voicevox or AivisSpeech application.
• Make sure that Dify containers are running with Docker.

Automatic setup (Recommended)

# Add executable permission
chmod +x install_voicevox_to_dify.sh

# Run the install script
./install_voicevox_to_dify.sh

# Run the install script with API endpoint
./install_voicevox_to_dify.sh http://localhost:50021/speakers

The script automatically performs the following steps:

• Clones the dify-voicevox-tts repository.
• Generates the voice list file (voicevox.yaml) for dify-voicevox-tts based on the speakers information installed on Voicevox or AivisSpeech.
• Overwrites the existing voicevox.yaml file.
• Installs the required modules (e.g., ffmpeg) in the docker-api-1 container.
• Installs dify-voicevox-tts in the docker-api-1 container.
• Commits the changes to the docker-api-1 container.
• Restarts the docker-api-1 container.

Manual setup

Generate yaml

# With Voicevox API
python generate_voicebox_yaml.py http://localhost:50021/speakers

## With AivisSpeech API
python generate_voicebox_yaml.py http://localhost:10101/speakers

Copy the yaml file to dify directory

Replace the generated yaml with the file in the dify-voicevox-tts directory.

Links

dify
dify-voicevox-tts
VOICEVOX
AivisSpeech