breeze-tts.json

{
  "_comment": "Replace the GGUF path and voice-library folder. One tts entry supports instruction-only speech and reference-audio conditioning. See README.md for both requests.",
  "host": "127.0.0.1",
  "port": 8080,
  "backend": "cuda",
  "device": 0,
  "threads": 4,
  "lazy_load": true,
  "max_loaded_models": 1,
  "ui": true,
  "ui_management": false,
  "_comment_voice_dir": "Replace /PATH/TO/voice-library with a folder you create. Put narrator.wav and a plain-text prompt_text file in it; see README.md.",
  "voice_dir": "/PATH/TO/voice-library",
  "models": [
    {
      "id": "breeze",
      "family": "breeze_tts",
      "path": "/PATH/TO/breeze-tts-2-q8_0.gguf",
      "task": "tts",
      "mode": "offline",
      "default_request_options": {
        "instruct": "Speak clearly and naturally.",
        "seed": 123
      }
    }
  ]
}