Resilient to noise, reverberation, and works across a variety of accents. Outperforms competing solutions by wide margins.
Built upon proprietary deep learning technology. The natural fit for IoT and resource-constrained applications.
Listen for multiple commands simultaneously, with no additional resource usage.
# Create an instance of Porcupinehandle = pvporcupine.create(keyword_paths=['/PATH/TO/YOUR/WAKE_WORD_MODEL'])# Process audiowhile True:keyword_index = handle.process(get_next_audio_frame())if keyword_index >= 0:# Insert detection callback here