Using SpeechRecognizer

SpeechRecognizer turns spoken audio into text using Android's speech recognition service. Add it to a GameObject when you want voice commands, dictation, or any interaction that should react to what the player says.

The component works as a stream of frame events:

• partialResults contains temporary recognition text on frames where Android reports a partial hypothesis while the user is still speaking.
• results contains final recognition text for exactly one frame, and only when Android closes a recognition pass.
• On frames with no new recognition data, both values are empty.
• Check the component every update and copy non-empty values to your own variable, UI text, command parser, or log when you need to keep them. If you do not read results on that completion frame, the final text is gone on the next frame.

Setup

1. Add SpeechRecognizer to a GameObject.
2. Set language to a BCP-47 tag such as pt-BR or en-US.
3. Enable enableRecognition when you want the component to listen.
4. Read partialResults and results during update.

On Android, the app needs microphone permission. If permission has not been granted, the component asks for it when recognition is enabled.

Example With Virtual Attributes

This version uses virtual attributes, which are the shortest way to work with the component in scripts.

SpatialObject myObject = /* your object */ null;
SpeechRecognizer speech = myObject.findComponent(SpeechRecognizer.class);
String lastFinalText = "";

if (speech != null) {
    speech.language = "pt-BR";
    speech.enableRecognition = true;
}

// Run this part every update.
if (speech != null) {
    String partialText = speech.partialResults;
    if (partialText != null && !partialText.isEmpty()) {
        Terminal.log("Partial speech: " + partialText);
    }

    String finalText = speech.results;
    if (finalText != null && !finalText.isEmpty()) {
        lastFinalText = finalText;
        Terminal.log("Final speech: " + lastFinalText);
    }
}

Example With Methods

Use the methods directly when you prefer explicit calls or when you want to make the getter/setter behavior obvious.

SpatialObject myObject = /* your object */ null;
SpeechRecognizer speech = myObject.findComponent(SpeechRecognizer.class);

if (speech != null) {
    speech.setLanguage("pt-BR");
    speech.setEnableRecognition(true);
}

// Run this part every update.
if (speech != null && speech.isEnableRecognition()) {
    String partialText = speech.getPartialResults();
    if (partialText != null && !partialText.isEmpty()) {
        Terminal.log("Partial speech: " + partialText);
    }

    String finalText = speech.getResults();
    if (finalText != null && !finalText.isEmpty()) {
        Terminal.log("Final speech: " + finalText);
    }
}

Notes

• Partial text may change as Android improves its hypothesis.
• Final text appears for exactly one frame when speech recognition finishes and is the best value to use for commands that should execute only once.
• If language is invalid, the component logs an error in the Terminal and disables recognition.
• Speech recognition depends on the Android recognition service available on the device.