Realtime Whisper

Type: nodetool.audio.RealtimeWhisper

Namespace: nodetool.audio

Description

Stream audio input to WhisperLive and emit real-time transcription. realtime, whisper, transcription, streaming, audio-to-text, speech-to-text

Emits:
  - `chunk` Chunk(content=..., done=False) for transcript deltas
  - `chunk` Chunk(content="", done=True) to mark segment end
  - `text` final aggregated transcript when input ends

Properties

Property Type Description Default
model Enum['tiny', 'base', 'small', 'medium', 'large', 'large-v2', 'large-v3'] Whisper model size - larger models are more accurate but slower tiny
language Enum['auto', 'en', 'es', 'fr', 'de', 'it', 'pt', 'nl', 'ru', 'zh', 'ja', 'ko', 'ar', 'hi', 'tr', 'pl', 'uk', 'vi'] Language code for transcription, or ‘auto’ for automatic detection en
chunk chunk The audio chunk to transcribe {'type': 'chunk', 'node_id': None, 'content_type': 'text', 'content': '', 'content_metadata': {}, 'done': False}
temperature float Sampling temperature for transcription 0.0
initial_prompt str Optional initial prompt to guide transcription style ``

Outputs

Output Type Description
start float  
end float  
text str  
chunk chunk  
speaker int  
detected_language str  
translation str  

Metadata

Browse other nodes in the nodetool.audio namespace.