#153 · Primary category: Speech & Audio

mycroft-precise

embedded-systems hotword-detection keyword-spotting raspberry-pi speech-recognition voice-control voice-recognition wake-word-detection

A lightweight, simple-to-use, RNN wake word listener

Project last updated:11/25/23

GitHub Stars

964

Forks

242

Contributors

13

License

Apache-2.0

Why we included this project

A voice assistant that listens continuously needs a cheap way to catch its wake phrase without decoding every word around it. Precise does exactly that: it runs a small recurrent network over the microphone stream and fires an event only when the trained phrase shows up. That keeps it light enough for a Raspberry Pi and easy to drop into your own application. You can train a trigger for anything from a product name to a cough, or start with the official and community models to avoid the data work. Embedded assistants and home automation projects get a focused component instead of a heavy framework to build around.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category