Nodes/ComfyUI Kaola MOSS-TTS/Generates MOSS Voice from Text
ComfyUI Node

Generates MOSS Voice from Text

Generates a voice sample based on a text description. Parameters: - instruction: Describe the voice (e.g., 'A young female, soft voice'). - text: The content to be spoken. - audio_temperature: 0.1-2.0. Higher = more variation. - audio_top_p/top_k: Sampling parameters.

By kana112233·Created 6 months ago·Updated 6 months ago· 3
Generates MOSS Voice from Text
  • moss_vg_model
  • audio
textHello, this is a test.
instructionA clear, neutral voice for reading.
audio_temperature1.0
audio_top_p0.80
audio_top_k50
audio_repetition_penalty1.0
max_new_tokens2000
text_normalizetrue
CategoryKaola/MOSS-TTSD

Inputs (9)

NameTypeDefaultDescription
moss_vg_modelMOSS_VOICE_GENERATOR_MODELLoaded MOSS-VoiceGenerator model.
textSTRINGHello, this is a test.The content to be spoken.
instructionSTRINGA clear, neutral voice for reading.Describe the desired voice characteristics (gender, age, tone, emotion).
audio_temperatureFLOAT1.00.1–2Higher = more variation/drama.
audio_top_pFLOAT0.800.1–1Nucleus sampling probability.
audio_top_kINT501–200Top-K sampling.
audio_repetition_penaltyFLOAT1.01–2Penalty for repeating audio tokens.
max_new_tokensINT2000100–10000Max audio length in tokens.
text_normalizeBOOLEANtrueNormalize text input.

Outputs (1)

NameTypeDescription
audioAUDIO