Class STT

Buffers voiced audio frames and converts them to text transcripts.

Index

Constructors

Methods

feed finalize onFinal onPartial reset

Constructors

constructor

new STT(config: SttConfig): STT
Creates a new STT instance.
Parameters
- config: SttConfig
  Language and audio format settings.
Returns STT
- Defined in apps/tutor/src/voice/STT.ts:97

Methods

feed

feed(chunk: Float32Array): void
Feeds a PCM chunk into the internal audio buffer. Should only be called while the VAD reports isSpeaking() === true.
Parameters
- chunk: Float32Array
  Float32 PCM samples at the configured sample rate.
Returns void
- Defined in apps/tutor/src/voice/STT.ts:107

finalize

finalize(): Promise<SttResult>
Signals end of the current utterance and triggers recognition. Clears the internal buffer after dispatching the result.

Returns Promise<SttResult>
The completed SttResult for the buffered utterance.
- Defined in apps/tutor/src/voice/STT.ts:117

onFinal

onFinal(cb: FinalTranscriptCallback): void
Registers a callback for the final SttResult.
Parameters
- cb: FinalTranscriptCallback
  Invoked once per completed utterance.
Returns void
- Defined in apps/tutor/src/voice/STT.ts:135

onPartial

onPartial(cb: PartialTranscriptCallback): void
Registers a callback for partial (streaming) transcript updates.
Parameters
- cb: PartialTranscriptCallback
  Invoked with partial text as it arrives from the backend.
Returns void
- Defined in apps/tutor/src/voice/STT.ts:126

reset

reset(): void
Discards any buffered audio without running recognition. Used when a session is aborted mid-utterance.

Returns void
- Defined in apps/tutor/src/voice/STT.ts:143