Config

Note

The config Function is non-blocking.

{
  "verb": "config",
  "synthesizer": {
    "voice": "Jenny"
  },
  "recognizer": {
    "vendor": "google",
    "language": "de-DE"
  },
  "bargeIn": {
    "enable": true,
    "input" : ["speech"],
    "actionHook: "/userInput"
    }
}

Parameters

enable answering machine detection; see the amd properties of the dial Function for details.

An object describing behaviors to apply if the application delays in responding to an action hook (same options as the gather Function).

this object contains properties that are used to instantiate a 'background' gather Function.

if true, begin listening for speech or dtmf input while the session is executing other Functions. This is known as a "background gather" and an application to capture user input outside of a gather Function. If false, stop any background listening task that is in progress.

If true and bargeIn.enable is true, then when the background gather completes with speech or dtmf detected, it will automatically start another background gather.

A webhook to call if user input is collected from the background gather.

Array, specifying allowed types of input: ['digits'], ['speech'], or ['digits', 'speech'].

Dtmf key that signals the end of dtmf input.

Exact number of dtmf digits expected to gather.

Minimum number of dtmf digits expected to gather. Defaults to 1.

Maximum number of dtmf digits expected to gather.

Amount of time to wait between digits after minDigits have been entered.

A string or integer value indicating the number of decibels to boost or reduce the strength of the outgoing audio signal to the caller/called party, e.g. "-6 dB". Note this applies to the main track only, not to any dub tracks.

play audio to the caller while the remote application is processing gather transcriptions. This is a session-wide setting that may be overridden at the gather Function level.

whether to enable or disable filler noise.

http(s) audio to play as filler noise.

integer value specifying number of seconds to wait for a response from the remote application before playing filler noise (default: play immediately after sending results).

a nested listen action, which allows recording of the call from this point forward by streaming the audio to a remote server over a websocket connection.

whether to enable event notifications (Function:status messages) over websocket connections. Functions that are sent over the websocket must also contain an "id" property to activate this feature.

provides the URL to a remote music source to use when a call is placed on hold.

change the session-level default speech recognition settings. See the transcribe Function for details on the recognizer property.

resets either 'recognizer' and/or 'synthesizer' to the default application settings.

options to manage call recording.

"startCallRecording", "stopCallRecording", "pauseCallRecording", or "resumeCallRecording".

Specifies the recording type.
Allowed values:

• siprec (default)
• cloud

When action : pauseCallRecording setting this to true will pad the recording with silence. Defaults to false meaning the recording file is truncated during the pause.

SIP URI for SIPREC server.
Note: Required if action is "startCallRecording" and type is not "cloud".

User-supplied string to identify the recording. Note: Only used with siprec recording

a nested transcribe action, which allows a transcription of the call to be sent in the background.

if true start the transcribe, if false stop it.

the webhook/websocket identifier to send transcriptions to (required if enabling transcription).

recognizer options.

a webhook to call when a SIP request is received within the dialog (e.g. an INFO, NOTIFY, or REFER).

change the session-level default text-to-speech settings. See the say Function for details on the synthesizer property.

disable TTS caching for an entire call session

Enable or disable server-side noise isolation to reduce background noise on call audio. Useful for improving STT accuracy in noisy environments.

Whether to enable or disable noise isolation.

Noise isolation vendor: "krisp" or "rnnoise". Availability of Krisp depends on your service plan. RNNoise is open-source and requires no API key.

Suppression level 0–100. Higher values are more aggressive. Default: 100.

Optional model name override.

Enable or disable turn-taking detection for conversational AI applications. Turn-taking uses an acoustic model to determine when the user has finished speaking, rather than relying on silence detection alone.

Whether to enable or disable turn-taking detection.

Turn-taking vendor: "krisp". Availability of Krisp depends on your service plan.

Confidence threshold 0.0–1.0. Lower values trigger earlier turn transitions (more aggressive). Default: 0.5.

Optional model name override.