Develop

Voice SDK activation

Updated: Apr 15, 2026
There are a few important methods for activation and deactivation of a Voice SDK voice experience. These methods are called on an AppVoiceExperience actor placed in your level. Each method is available as both a Blueprint node and a C++ function (exposed via UFUNCTION(BlueprintCallable)). For details on setting up the actor in your map, see Voice Command.
Note: An end user has the choice to consent to mic use in the developer app at the start of the app use. Only if the end user turns on the mic activation will the developer app record the voice command and send it directly to Wit.ai for Natural Language Understanding (NLU).

Prerequisites

Before using these activation methods, integrate the Voice SDK plugin into your project and configure a Wit.ai app. For setup instructions, see Integrating Voice SDK.

Activation methods

  • SendTranscription(string) - This method takes the content of a provided string and sends it to Wit.ai for the NLU to process. This is useful if you are using some form of on-device Automatic Speech Recognition (ASR) for transcriptions.
  • ActivateVoiceInput() - This method turns on the default mic on the end user’s device and listens for voice commands after the volume threshold is hit. By default, the mic records for up to 20 seconds and stops if the user is quiet for 2 seconds.
  • ActivateVoiceInputImmediately() - This method immediately turns on the default mic on the end user’s device (if not already on) and begins transmitting the data to Wit.ai for the NLU to process. Unlike ActivateVoiceInput(), this method skips the wake volume threshold and starts streaming immediately.
  • DeactivateVoiceInput() - This method stops the microphone from recording and sends any data collected. This enables you to have more control over your activation. For example, an application could activate recording when the user presses the grip button and deactivate when they release it.
The recording duration and silence timeout for ActivateVoiceInput() can be adjusted via the MaximumRecordingTime and KeepAliveTime values in the Voice Configuration settings in the Unreal Editor.
Nav Logo
Virtual reality / Meta Horizon
Virtual reality / Meta Horizon
Virtual reality / Meta Horizon
English (US)
© 2026 Meta