. Browser-Speech: Simplified HTML5 Speech Synthesis

Rick Carlino

Personal blog of Rick Carlino, senior software engineer at Qualia Labs, co-founder of Fox.Build Makerspace. Former co-founder of FarmBot.

Browser-Speech: Simplified HTML5 Speech Synthesis

UPDATE (2020): Google has blocked text-to-speech autoplay, thereby rendering this NPM module only useful for Firefox users under most circumstances. Sorry.

I published a new NPM module to simplify access to the HTML5 Text To Speech API.

If you are inpatient and just want to see what it does, you can take a look at the live demo page.

The download page for the NPM package is here

What Is It?

The HTML5 Text To Speech API is a powerful tool for making web pages that talk.

The API comes with a large number of configuration options and callbacks. Such settings are necessary for special use cases. For most of my TTS use cases, I’ve found I really only need a few things:

  • An input string
  • A language code (Eg: en, es, ko, etc..)
  • A callback that fires on start
  • An error callback

browser-speech is born from these needs. The library is 0 config by design and only does one thing: talk.

If you don’t need a bunch of knobs and special customization for your speech use cases, this library may be a good fit.

If you do require special settings (such as speech rate adjustment, pauses, etc…) using the HTML5 Speech API directly may be a better choice.

browser-speech is written in Typescript and targets all modern browsers.

Is it any good?

Yes. It is used in real world applications, such as the control software for FarmBot. If you would like your application added to a list of real-world use cases, please submit a PR.


npm install browser-speech


The library exports a talk(text: string, language: string) function. The function is promise based and works well with the await keyword.

import { talk } from "browser-speech";

// FIRST PARAMETER:  Text you wish to speak
// SECOND PARAMETER: Language code. Default is "en" (English). Throws exception
                     when an unsupported code is entered.
talk("Hello, world!", "en")
  .then((event) => console.log("Text to speech is nice");)
  .catch((event) => console.error("Failed. :("););

Running Live Demo

Load integration_test.html in your browser via parcel.

npm install -g parcel-bundler
parcel live_demo.html
# You can now test the operability of the `talk()` function by opening
# http://localhost:1234 - You should hear speech.