Nodes/ComfyUI-OllamaOmni/Ollama Audio Transcribe
ComfyUI Node

Ollama Audio Transcribe

Transcribe audio to text using an Ollama speech-to-text model such as whisper.

By ckinpdx·Created 4 months ago·Updated 4 months ago· 1
Ollama Audio Transcribe
  • audio
  • text
urlhttp://127.0.0.1:11434
model
prompt
language
temperature0.00
CategoryOllama/Audio

Inputs (6)

NameTypeDefaultDescription
audioAUDIOAudio to transcribe. Connect from Load Audio or any AUDIO output.
urlSTRINGhttp://127.0.0.1:11434URL of the Ollama server.
modelCOMBOSpeech-to-text model (e.g. whisper, whisper-large-v3-turbo).
promptoptSTRINGOptional hint to guide transcription (vocabulary hints, speaker style, etc.).
languageoptSTRINGBCP-47 language code (e.g. 'en', 'fr', 'ja'). Leave blank for auto-detect.
temperatureoptFLOAT0.000–1Sampling temperature. 0 = deterministic.

Outputs (1)

NameTypeDescription
textSTRING