Skip to content

TranscriptionSpeechFinalEvent

Emitted when an utterance is fully complete.

Defined in: src/core/events/types.ts:194

Emitted when an utterance is fully complete.

Remarks

This is the canonical trigger for LLM processing. For Deepgram, this fires when speech_final=true (the speaker has stopped talking). For NativeSTT and other providers that emit one result per utterance, this equals transcription.final.

Multi-segment providers (Deepgram) may emit several transcription.final events for a single utterance. Only the last one is followed by transcription.speechFinal.

Example

agent.on('transcription.speechFinal', (event) => {
  console.log('Complete utterance:', event.text);
});

See

Extends

  • BaseEvent

Properties

PropertyTypeDescriptionInherited fromDefined in
confidence?numberConfidence score (0-1) for this result, if available.-src/core/events/types.ts:202
metadata?Record<string, unknown>Optional metadata associated with the event. Remarks May contain provider-specific data or debugging information.BaseEvent.metadatasrc/core/events/types.ts:88
textstringThe complete transcribed text for the entire utterance.-src/core/events/types.ts:199
timestampnumberUnix timestamp (in milliseconds) when the event occurred. Remarks Useful for latency measurements and debugging the pipeline timing.BaseEvent.timestampsrc/core/events/types.ts:80
type"transcription.speechFinal"Discriminant for this event type.-src/core/events/types.ts:196

© 2026 CompositeVoice. All rights reserved.

Font size
Contrast
Motion
Transparency