bevy_whisperer

Speech recognition plugin for bevy, using whisper-apr.

The microphone feature enables speech transcription from a microphone.

You need an .apr model to use the plugin. You can download the multilingual whisper-tiny model here.

curl -L -o assets/whisper-tiny.apr https://huggingface.co/paiml/whisper-apr-tiny/resolve/main/whisper-tiny.apr

You can also generate your own apr models, by converting existing Whisper models to the apr model format. If you only need the English language, you can convert the tiny.en Whisper model to the apr format. See this page for a list of the different models.

Examples

Asset

The transcribe_asset example shows how to run speech recognition on a bevy asset.

Control

The control example shows how to use the WhispererMicPlugin and control an entity by listening to transcription events.

Microphone

The old transcribe_microphone example shows how to manually run speech recognition on the audio coming from a microphone.

Compatible Bevy versions