Skip to content

Quick Start

Three ways to get speech-to-text running, from simplest to most flexible.

Option A: Drop-in Actor (Fastest)

  1. Download a Whisper model via Project Settings > Aboleth Speech-to-Text > Model Management.
  2. Drag AbolethSTTListenerActor into your level.
  3. In the Details panel, bind OnUtteranceProcessed to receive transcribed text.
  4. Play. Talk. Text appears.

No Blueprint wiring required. The actor auto-loads the Whisper model, opens the microphone, and starts listening on BeginPlay.

Option B: Listener Component

  1. Open any Blueprint actor (e.g. your PlayerCharacter).
  2. Add Component > Aboleth STT Listener.
  3. Bind OnUtteranceProcessed on the component.
  4. Play.

Same behavior as the actor, but attached to an existing actor instead of standalone.

Option C: Subsystem Direct (Full Control)

// Get the GameInstanceSubsystem
UAbolethSTTSubsystem* STT = UAbolethSTTSubsystem::GetSTTSubsystem(this);

// Load Whisper model + Silero VAD
STT->LoadSTTSystem();

// Start microphone and listening
STT->StartListening();

// Bind the transcription event
STT->OnUtteranceProcessed.AddDynamic(this, &AMyActor::OnSpeechRecognized);

void AMyActor::OnSpeechRecognized(const FString& Text)
{
    UE_LOG(LogTemp, Log, TEXT("Player said: %s"), *Text);
}

Use the built-in Get Subsystem node (GameInstance > AbolethSTTSubsystem) to access the subsystem, then call Load STT System, Start Listening, and bind On Utterance Processed.

What's Next?