@intentface/react-speech-recognition

@intentface/react-speech-recognition is a custom React hook that provides an easy-to-use interface for integrating speech recognition capabilities into your React applications. It leverages the native browser SpeechRecognition API, allowing you to capture and process voice inputs with customizable options.

Installation

You can install this package via npm:

1npm install @intentface/react-speech-recognition

Usage

Here’s a basic example of how to use the useSpeechRecognition hook in your React project:

1import React from "react";
2import { useSpeechRecognition } from "@intentface/react-speech-recognition";
3
4const SpeechComponent = () => {
5  const {
6    transcript,
7    interimTranscript,
8    isListening,
9    isFinal,
10    isSupported,
11    start,
12    stop,
13    error,
14  } = useSpeechRecognition({
15    lang: "en-US", // Language for speech recognition (default is "en-US")
16    continuous: false, // Whether to keep listening or stop after receiving input
17    timeout: 5000, // Automatically stop listening after 5 seconds
18    onUpdate: ({ transcript, interimTranscript, isFinal }) => {
19      console.log("Update:", { transcript, interimTranscript, isFinal });
20    },
21    onError: ({ error }) => {
22      console.error("Speech recognition error:", error);
23    },
24  });
25
26  if (!isSupported) {
27    return <p>Your browser does not support Speech Recognition.</p>;
28  }
29
30  return (
31    <div>
32      <h1>Speech Recognition Example</h1>
33      <p>Transcript: {transcript}</p>
34      <p>Interim Transcript: {interimTranscript}</p>
35      <p>
36        Status:{" "}
37        {isListening ? "Listening..." : isFinal ? "Finalized" : "Stopped"}
38      </p>
39      {error && <p>Error: {error.message}</p>}
40      <button onClick={start}>Start Listening</button>
41      <button onClick={stop}>Stop Listening</button>
42    </div>
43  );
44};
45
46export default SpeechComponent;

API

The useSpeechRecognition hook returns an object containing the following properties:

Returned Values

• transcript: The final processed transcript of the recognized speech.
• interimTranscript: The current intermediate transcript while speech is being processed.
• isListening: A boolean indicating whether the recognition is currently active.
• isFinal: A boolean indicating whether the current session has ended and the transcript is finalized.
• isSupported: A boolean indicating whether the browser supports speech recognition.
• start(): A function to start speech recognition.
• stop(): A function to stop speech recognition.
• error: An object containing any speech recognition errors that occurred.

Options

The hook accepts an options object with the following fields:

• lang: The language for speech recognition (default is "en-US").
• continuous: If true, the recognition will continue until stopped manually (default: false).
• timeout: Time in milliseconds to automatically stop recognition after no speech is detected (default: undefined). If left undefined, browsers default behaviour is respected.
• onUpdate: A callback function triggered when there is an update in the transcript. Receives an object with { transcript, interimTranscript, isFinal }.
• onError: A callback function triggered when an error occurs. Receives an object with { error }.

Features

• Customizable Language: Set the language of recognition with the lang option.
• Continuous Listening: Option to listen continuously for speech input.
• Timeout Control: Automatically stop recognition after a set period of inactivity.
• Error Handling: Receive error information via the onError callback.
• Interim and Final Transcripts: Access both the interim (in-progress) and final transcripts.

Browser Compatibility

This hook uses the native SpeechRecognition API (also known as webkitSpeechRecognition in some browsers). Below is a list of supported browsers:

Browser	Supported Versions
Chrome	Version 33+
Safari	Version 14+
Edge	Version 79+

Note: The SpeechRecognition API is not supported in Firefox or Internet Explorer. Please check for browser compatibility using the isSupported flag provided by the hook.

Contributing

Contributions are welcome! If you encounter any issues or have suggestions for improvements, feel free to create a pull request or file an issue in the repository.

Please remember to add the changes you are making by running

1npm run add-change

License

MIT

Dev

Doing a release

After finishing your feature normally, to automatically bump version number, update changelog and do an npm release:

1. npm run add-change
2. Follow the instructions
3. Commit the changeset file(s) to your feature branch and push to Github
4. Create a PR normally, including the changeset file(s)
5. Eventually accept the PR called "Version Packages" created by the changeset bot

About us

For more details about who we are, visit our website: Intentface.