This demo uses Porcupine for Web and the WebVoiceProcessor to:
Create an English instance of Porcupine that listens for "Picovoice" (a
built-in keyword, provided by name) and "Papaya Whip" (a custom keyword
provided via its base64 representation).
Create a German instance of Porcupine to listen for "Ananas".
Create a French instance of Porcupine to listen for "Framboise".
Create a Spanish instance of Porcupine to listen for "Manzana".
Acquire microphone (& ask permission) data stream and convert to voice
processing format (16kHz 16-bit linear PCM). The downsampled audio is
forwarded to the Porcupine engines. The audio does not leave the
browser: all processing is occurring via the Porcupine WebAssembly code.
Listen for keyword detection events from the Porcupine engines and
output them to the page.
After entering the AccessKey, click the "Start Porcupine" button.