ComfyUI Node
BlueMagpie TTS
Generate 48 kHz Taiwanese Mandarin or Mandarin/English speech with a built-in or rights-cleared reference voice.
BlueMagpie TTS
- model
- reference_audio
- audio
- generation_info
◄text你好，這是 BlueMagpie TTS 的 ComfyUI 測試。►
◄speakerhung_yi_lee►
◄cfg_value2.0►
◄inference_timesteps10►
◄max_len2000►
◄retry_badcasetrue►
◄reference_audio_path►
Categoryaudio/BlueMagpie TTS
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| model | BLUE_MAGPIE_MODEL | — | |
| text | STRING | 你好，這是 BlueMagpie TTS 的 ComfyUI 測試。 | — |
| speaker | COMBO | hung_yi_lee | 2 options: hung_yi_lee, female_voice |
| cfg_value | FLOAT | 2.01–4 | — |
| inference_timesteps | INT | 101–50 | — |
| max_len | INT | 200016–8192 | — |
| retry_badcase | BOOLEAN | true | — |
| reference_audioopt | AUDIO | Optional rights-cleared voice reference. Use one clean speaker for about 3 to 10 seconds; it overrides speaker selection. | |
| reference_audio_pathopt | STRING | Optional local reference audio path (for example WAV, FLAC, or MP3). Use a clean 3 to 10 second clip. Ignored when reference_audio is connected. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |
| generation_info | STRING | — |