Nodes/MOSS-TTS/MOSS-TTS Whisper Transcribe
ComfyUI Node

MOSS-TTS Whisper Transcribe

Transcribe AUDIO with Whisper for MOSS-TTS continuation prefix text.

By Saganaki22·Created 2 months ago·Updated 2 months ago· 31
MOSS-TTS Whisper Transcribe
  • audio
  • transcript
modelwhisper-large-v3-turbo (auto-download)
dtypeauto
languageauto
tasktranscribe
chunk_length_s30
download_if_missingtrue
CategoryMOSS-TTS

Inputs (7)

NameTypeDefaultDescription
audioAUDIOAudio to transcribe. Useful for Continue Speech prefix_text.
modelCOMBOwhisper-large-v3-turbo (auto-download)Whisper ASR model. Turbo is fast and usually accurate enough for prefix transcripts.
dtypeCOMBOautoWhisper precision. auto uses bf16 on supported CUDA/XPU and fp32 otherwise.
languageCOMBOautoAudio language. auto detects it; setting it can improve transcript accuracy.
taskCOMBOtranscribetranscribe keeps the original language; translate outputs English.
chunk_length_sINT300–120Whisper chunk length for longer clips. 0 lets Transformers choose.
download_if_missingBOOLEANtrueDownload the selected Whisper model into ComfyUI/models/audio_encoders if it is missing.

Outputs (1)

NameTypeDescription
transcriptSTRING