Rhino - NodeJS API

Speech-to-Intent Engine
Domain Specific NLU
Offline NLU
Local Voice Recognition
Raspberry Pi
Linux
macOS
NodeJS
NPM

This document outlines how to integrate the Rhino Speech-to-Intent engine within an NodeJS application.

Requirements

NodeJS 10+
yarn (or npm)

Compatibility

Linux (x86_64)
macOS (x86_64)
Raspberry Pi (2,3,4)

Web Browsers

Looking to run Picovoice in-browser? See the JavaScript WebAssembly demo instead.

Installation

yarn add @picovoice/rhino-node

npm install @picovoice/rhino-node

Usage

The Rhino SDK for NodeJS provides the Rhino class. Here is an example using the sample "Coffee Maker" context file:

const Rhino = require("@picovoice/rhino-node")
 
const coffeeMakerContextPath = "./coffee_maker.rhn"
 
let handle = new Rhino(coffeeMakerContextPath)
 
let isFinalized = false
// process each frame of audio until Rhino has concluded that it understood the phrase (or did not)
// when Rhino has reached a conclusion, isFinalized will become true
while (!isFinalized) {
  isFinalized = handle.process(frame)
  // retrieve the inference from Rhino
  if (isFinalized) {
    let inference = handle.getInference()
    // inference result example:
    //
    //   {
    //     isUnderstood: true,
    //     intent: 'orderDrink',
    //     slots: {
    //       size: 'medium',
    //       numberOfShots: 'double shot',
    //       coffeeDrink: 'americano',
    //       milkAmount: 'lots of milk',
    //       sugarAmount: 'some sugar'
    //     }
    //   }
  }
}
 
// always call release when finished to free the resources allocated by Rhino
handle.release()

Override model and library paths

The Rhino constructor accepts three optional positional parameters for the sensitivity and the absolute paths to the model and dynamic library, should you need to override them (typically, you will not).

let handle = new Rhino(contextPath, sensitivity, modelFilePath, libraryFilePath)

Custom Context

You can create custom Rhino context models using Picovoice Console.