VOICEVOX CLI

Japanese text-to-speech CLI using VOICEVOX Core for Apple Silicon Macs (requires Nix).

Quick Start

nix shell github:usabarashi/voicevox-cli

voicevox-setup                              # Download resources (first time only)
voicevox-say "こんにちは、ずんだもんなのだ"   # Voice synthesis

Usage

# Voice synthesis (daemon starts automatically)
voicevox-say "こんにちは、ずんだもんなのだ"
voicevox-say --speaker-id 3 "声を変えてみるのだ"
voicevox-say -o output.wav "保存するテキスト"
echo "パイプからの入力" | voicevox-say

# Voice discovery
voicevox-say --list-speakers
voicevox-say --status

# Daemon management
voicevox-daemon --start
voicevox-daemon --stop
voicevox-daemon --restart
voicevox-daemon --status

MCP Server

Enable AI assistants to use VOICEVOX for Japanese speech synthesis.

voicevox-mcp-server

See detailed MCP documentation

Usage Notes

• If you customize VOICEVOX_SOCKET_PATH, use a directory owned by the same user with restrictive permissions (recommended: 0700 for the directory).
• In shared environments, avoid placing the socket in a world-writable directory.

Troubleshooting

voicevox-say --status              # Check installation status
voicevox-setup                     # Reinstall all resources
voicevox-setup --purge             # Remove all local data for a clean reinstall
voicevox-daemon --restart          # Restart daemon
GH_TOKEN=$(gh auth token) voicevox-setup  # Avoid GitHub API rate limits

License

See LICENSE for details. Generated audio requires credit "VOICEVOX:[Character Name]" (e.g., "VOICEVOX:ずんだもん"). License terms are displayed during voicevox-setup.

Details: VOICEVOX Terms of Use

---

ずんだもんと一緒に楽しい TTS ライフを送るのだ！