Porcupine
A highly accurate, lightweight, on-device wake-word engine for building always-listening voice-enabled applications.
Porcupine uses deep neural networks trained in real-world environments, stays compact and computationally efficient for IoT use, and runs across Arm Cortex-M, STM32 and Arduino boards, Raspberry Pi, Android and iOS, Chrome, Safari, Firefox and Edge, and Linux, macOS and Windows. It can detect multiple always-listening voice commands with no added runtime footprint, and a developer can train custom wake-word models with Picovoice Console.
The material reports Porcupine as 11.0 times more accurate and 6.5 times faster than the better-performing of two other engines it was compared against, PocketSphinx and Snowboy, on a Raspberry Pi 3. It supports English, Mandarin Chinese, French, German, Italian, Japanese, Korean, Portuguese and Spanish, with more languages available to commercial customers.
It ships SDKs for Python, .NET, Java, Flutter, React Native, Android, iOS, web and C; it is licensed under Apache-2.0.