Skip to main content

SileroVAD

Silero voice-activity detector (VAD).

Constructor

threshold
number
Speech-probability threshold to start speech, in [0, 1]. Default 0.4.
startThreshold
number | null
Overrides threshold when set. Default null.
endThreshold
number | null
Overrides stopThreshold when set. Default null.
stopThreshold
number
Probability below which speech ends, in [0, 1]. Default 0.25.
minSpeechDuration
number
Minimum speech duration in seconds. Default 0.3.
minSilenceDuration
number
Minimum trailing silence to end a turn, in seconds. Default 0.4.
paddingDuration
number
Audio padding around detected speech, in seconds. Default 0.5.
maxBufferedSpeech
number
Maximum buffered speech, in seconds. Default 60.0.
sampleRate
number
Model sample rate in Hz (used unless modelSampleRate is set). Default 16000.
inputSampleRate
number | null
Sample rate (Hz) of the incoming audio. Default null (treated as 48000).
modelSampleRate
number | null
Overrides sampleRate as the model sample rate. Default null.
smoothingFactor
number
Exponential smoothing factor for speech probability. Default 0.35.
forceCpu
boolean
Force CPU inference. Default false.
minVolume
number
Minimum input volume gate; ignored unless energyFilterEnabled. Default 0.0.
energyFilterEnabled
boolean
Enable the energy/volume gate. Default true. When false, minVolume is forced to 0.0.

close

Release any resources held by the provider.

closeEmitter

Close the emitter: drop all handlers and stop emitting. Waits up to 2 seconds for in-flight async handlers to settle before returning.

emit

event
string
required

flush

Flush any buffered audio.

getRuntimeConfig

returns
Record<string, any>

listenerCount

event
string
required
returns
number

off

event
string
required
handler
EventHandler
required

on

event
string
required
handler
EventHandler
required
returns
EventHandler

once

event
string
required
handler
EventHandler
required
returns
EventHandler

onVadEvent

Register a callback to receive VADResponse events.
callback
VadEventCallback
required

processAudio

Process a buffer of audio frames. Override in a custom provider.

removeAllListeners

event
string