Picovoice Console — Porcupine Wake Word Engine

Navigate to the Wake Word console from the Picovoice Console landing page:

Type in the phrase

A unique feature of Porcupine wake word engine is that one can simply type in the trigger phrase and it will train a model for that specific phrase within hours. Picovoice uses a technique called transfer learning, which mimics how humans learn new phrases. When in the wake word console, type the wake phrase you want to train:

The choice of phrase is important for both accuracy and user experience. A good wake phrase needs to have some properties in terms of length and number of syllables. See Tips for Choosing a Wake Word for a summary.

Select your platform

The wake word models created by Console are platform-specific. This means that a model trained for Raspberry Pi cannot run on Android, for example. Select the target platform:

Create a wake word draft

A wake word draft is the set of instructions needed to kick off a model training job. It encapsulates both the typed-in phrase, the selected platform, and phrase validation. Create a draft:

If the phrase is accepted (pronunciation is understood and phrase length is not too short or too long), the draft appears:

Train a model from the draft

Click on the "Train" button, accept the terms of use, and click "Train" again to confirm:

The model is now being trained. You will be able to download it within a few hours.

Test the model inside your browser

A unique feature of Picovocie Console is that it allows testing within the browser. Test the newly-trained model in your browser:

Even if you turn off your WiFi, the testing will continue to function, as all processsing is performed in the browser itself.

Download the model

Click "Download" to get a copy of your wake word model (file with .ppn extension).

Use the .ppn file with the Picovoice SDK or Porcupine directly as a wake word / always-listening component in your voice user interface.