Picovoice SDK for Web

Picovoice

Made in Vancouver, Canada by Picovoice

Picovoice is an end-to-end platform for building voice products on your terms. It enables creating voice experiences similar to Alexa and Google. But it entirely runs 100% on-device. Picovoice is

• Private: Everything is processed offline. Intrinsically HIPAA and GDPR-compliant.
• Reliable: Runs without needing constant connectivity.
• Zero Latency: Edge-first architecture eliminates unpredictable network delay.
• Accurate: Resilient to noise and reverberation. It outperforms cloud-based alternatives by wide margins *.
• Cross-Platform: Design once, deploy anywhere. Build using familiar languages and frameworks.

Compatibility

• Chrome / Edge
• Firefox
• Safari

Restrictions

IndexedDB is required to use Picovoice in a worker thread. Browsers without IndexedDB support (i.e. Firefox Incognito Mode) should use Picovoice in the main thread.

Installation

Package

Using yarn:

yarn add @picovoice/picovoice-web

or using npm:

npm install --save @picovoice/picovoice-web

AccessKey

Picovoice requires a valid AccessKey at initialization. AccessKey acts as your credentials when using Picovoice SDKs. You can get your AccessKey for free. Make sure to keep your AccessKey secret. Signup or Login to Picovoice Console to get your AccessKey.

Usage

Picovoice requires a Porcupine keyword file (.ppn), a Rhino context file (.rhn) and model parameter files for both engines (.pv).

Each file offers two options on how to provide it to Picovoice:

Public Directory

NOTE: Due to modern browser limitations of using a file URL, this method does not work if used without hosting a server.

This method fetches the given file from the public directory and uses it to initialize Picovoice. Set the publicPath string to use this method.

Base64

NOTE: This method works without hosting a server, but increases the size of the model file roughly by 33%.

This method uses a base64 string of the given file and uses it to initialize Picovoice.

Use the built-in script pvbase64 to base64 your .ppn, .rhn or .pv file:

npx pvbase64 -i ${PICOVOICE_FILE} -o ${BASE64_FILENAME}.js

The output will be a js file containing a string which you can import into any file of your project. Set the base64 string with the imported js string use this method.

Picovoice Initialization Files

Picovoice saves and caches your model (.pv), keyword (.ppn) and context (.rhn) files in the IndexedDB to be used by Web Assembly. Use a different customWritePath variable choose the name the file will have in storage and set the forceWrite value to true to force an overwrite of the file. If the file changes, version should be incremented to force the cached file to be updated.

Either base64 or publicPath must be set for each file to instantiate Picovoice. If both are set for a particular file, Picovoice will use the base64 parameter.

// Custom keyword (.ppn)
const porcupineKeyword = {
  publicPath: ${KEYWORD_RELATIVE_PATH},
  // or
  base64: ${KEYWORD_BASE64_STRING},
  label: ${KEYWORD_LABEL},

  // Optional
  customWritePath: 'custom_keyword',
  forceWrite: true,
  version: 1,
  sensitivity: 0.6
}

// Context (.rhn)
const rhinoContext = {
  publicPath: ${CONTEXT_RELATIVE_PATH},
  // or
  base64: ${CONTEXT_BASE64_STRING},

  // Optionals
  customWritePath: 'custom_context',
  forceWrite: true,
  version: 1,
  sensitivity: 0.3,
}

// Model (.pv)
const porcupineOrRhinoModel = {
  publicPath: ${MODEL_RELATIVE_PATH},
  // or
  base64: ${MODEL_BASE64_STRING},

  // Optionals
  customWritePath: 'custom_model',
  forceWrite: true,
  version: 1,
}

Additional engine options are provided via the options parameter. Set processErrorCallback to handle errors if an error occurs while processing audio. Use endpointDurationSec and requireEndpoint to control the engine's endpointing behaviour. An endpoint is a chunk of silence at the end of an utterance that marks the end of spoken command.

// Optional. These are the default values
const options = {
  endpointDurationSec: 1.0,
  requireEndpoint: true,
  processErrorCallback: (error) => {},
}

Initialize Picovoice

Create wakeWordCallback and inferenceCallback functions to capture results from the engine:

function wakeWordCallback(wakeWordDetection) {
  console.log(`Picovoice detected keyword: ${wakeWordDetection.label}`);
}

function inferenceCallback(inference) {
  if (inference.isUnderstood) {
    console.log(inference.intent)
    console.log(inference.slots)
  }
}

Create an options object and add a processErrorCallback function if you would like to catch errors:

function processErrorCallback(error: string) {
...
}
options.processErrorCallback = processErrorCallback;

Initialize an instance of Picovoice in the main thread:

const picovoice = await Picovoice.create(
  ${ACCESS_KEY},
  porcupineKeyword,
  wakeWordCallback,
  porcupineModel,
  rhinoContext,
  inferenceCallback,
  rhinoModel,
  options // optional parameters
);

Or initialize an instance of Picovoice in a worker thread:

const picovoice = await PicovoiceWorker.create(
  ${ACCESS_KEY},
  porcupineKeyword,
  wakeWordCallback,
  porcupineModel,
  rhinoContext,
  inferenceCallback,
  rhinoModel,
  options // optional parameters
);

Process Audio Frames

Feed audio into the process() function. To start listening for the wake word and follow-on command. The result is received via wakeWordCallback and inferenceCallback as defined above.

function getAudioData(): Int16Array {
... // function to get audio data
  return new Int16Array();
}
for (; ;) {
  await picovoice.process(getAudioData());
  // break on some condition
}

Clean Up

Clean up used resources by Picovoice or PicovoiceWorker:

await picovoice.release();

Terminate (Worker Only)

Terminate PicovoiceWorker instance:

await picovoice.terminate();

Custom Keyword and Contexts

Create custom keywords and contexts using the Picovoice Console. Train a Porcupine keyword to obtain a keyword file (.ppn) and a Rhino context to obtain a context file (.rhn). To use them with the Web SDK, train the keywords and contexts for the target platform Web (WASM). These model files can be used directly with publicPath, but if base64 is preferable, convert to base64 JavaScript variable using the built-in pvbase64 script:

npx pvbase64 -i ${INPUT_BINARY_FILE}.{ppn/rhn} -o ${OUTPUT_BASE64_FILE}.js -n ${BASE64_VAR_NAME}

Similar to the model file (.pv), these files are saved in IndexedDB to be used by Web Assembly. Either base64 or publicPath must be set for each file to initialize Picovoice. If both are set, Picovoice will use the base64 model.

const picovoiceFile = {
  publicPath: "${FILE_RELATIVE_PATH}",
  // or
  base64: "${FILE_BASE64_STRING}",
}

Switching Languages

In order to use Picovoice with different languages you need to use the corresponding model file (.pv) for the desired language. The model files for all supported languages are available in the Porcupine and Rhino GitHub repositories.

Demo

For example usage refer to our Web demo application.