node package manager


Give your JavaScript the ability to speak many languages.


Build Status

Join the chat at

Polyglot.js is a tiny I18n helper library written in JavaScript, made to work both in the browser and in CommonJS environments (Node). It provides a simple solution for interpolation and pluralization, based off of Airbnb's experience adding I18n functionality to its Backbone.js and Node apps. Polyglot has zero dependencies.

I18n is incredibly important for us at Airbnb, as we have listings in 192 countries, and we translate our site into 30-odd different languages. We're also hiring talented engineers to help us scale up to meet the challenges of buliding a global marketplace.

View the documentation on Github.

View the annotated source.

Download the development version: 7.9kb, unminified with comments.

Download the production version: 2.8kb, minified (1.3kb gzipped).

Polylglot is agnostic to your translation backend. It doesn't perform any translation; it simply gives you a way to manage translated phrases from your client- or server-side JavaScript application.


For use with Node, install with NPM:

$ npm install node-polyglot

Running the tests

Clone the repo, run npm install, and npm test.



First, create an instance of the Polyglot class, which you will use for translation.

var polyglot = new Polyglot();

Polyglot is class-based so you can maintain different sets of phrases at the same time, possibly in different locales. This is very useful for example when serving requests with Express, because each request may have a different locale, and you don't want concurrent requests to clobber each other's phrases.


Tell Polyglot what to say by simply giving it a phrases object, where the key is the canonical name of the phrase and the value is the already-translated string.

  "hello": "Hello"
=> "Hello"

You can also pass a mapping at instantiation, using the key phrases:

var polyglot = new Polyglot({phrases: {"hello": "Hello"}});

Polyglot doesn't do the translation for you. It's up to you to give it the proper phrases for the user's locale.

A common pattern is to gather a hash of phrases in your backend, and output them in a <script> tag at the bottom of the document. For example, in Rails:


def index
  @phrases = { 
    "home.login" => I18n.t("home.login"),
    "home.signup" => I18n.t("home.signup"),


  var polyglot = new Polyglot({phrases: <%= raw @phrases.to_json %>});

And now you can utilize i.e. polyglot.t("home.login") in your JavaScript application or Handlebars templates.


Polyglot.t() also provides interpolation. Pass an object with key-value pairs of interpolation arguments as the second parameter.

  "hello_name": "Hola, %{name}."
polyglot.t("hello_name", {name: "DeNiro"});
=> "Hola, DeNiro."

Polyglot also supports nested phrase objects.

  "nav": {
    "hello": "Hello",
    "hello_name": "Hello, %{name}",
    "sidebar": {
      "welcome": "Welcome"
=> "Welcome"


For pluralization to work properly, you need to tell Polyglot what the current locale is. You can use polyglot.locale("fr") to set the locale to, for example, French. This method is also a getter:

=> "fr"

You can also pass this in during instantiation.

var polyglot = new Polyglot({locale: "fr"});

Currently, the only thing that Polyglot uses this locale setting for is pluralization.

Polyglot provides a very basic pattern for providing pluralization based on a single string that contains all plural forms for a given phrase. Because various languages have different nominal forms for zero, one, and multiple, and because the noun can be before or after the count, we have to be overly explicit about the possible phrases.

To get a pluralized phrase, still use polyglot.t() but use a specially-formatted phrase string that separates the plural forms by the delimeter ||||, or four vertical pipe characters.

For pluralizing "car" in English, Polyglot assumes you have a phrase of the form:

  "num_cars": "%{smart_count} car |||| %{smart_count} cars",

In English (and German, Spanish, Italian, and a few others) there are only two plural forms: singular and not-singular.

Some languages get a bit more complicated. In Czech, there are three separate forms: 1, 2 through 4, and 5 and up. Russian is even crazier.

polyglot.t() will choose the appropriate phrase based on the provided smart_count option, whose value is a number.

polyglot.t("num_cars", {smart_count: 0});
=> "0 cars"
polyglot.t("num_cars", {smart_count: 1});
=> "1 car"
polyglot.t("num_cars", {smart_count: 2});
=> "2 cars"

As a shortcut, you can also pass a number to the second parameter:

polyglot.t("num_cars", 2);
=> "2 cars"

Public Methods


Use extend to tell Polyglot how to translate a given key.

  "hello": "Hello",
  "hello_name": "Hello, %{name}"

The key can be any string. Feel free to call extend multiple times; it will override any phrases with the same key, but leave existing phrases untouched.


Clears all phrases. Useful for special cases, such as freeing up memory if you have lots of phrases but no longer need to perform any translation. Also used internally by replace.


Completely replace the existing phrases with a new set of phrases. Normally, just use extend to add more phrases, but under certain circumstances, you may want to make sure no old phrases are lying around.

Polyglot.prototype.t(key, interpolationOptions)

The most-used method. Provide a key, and t() will return the phrase.

=> "Hello"

The phrase value is provided first by a call to polyglot.extend() or polyglot.replace().

Pass in an object as the second argument to perform interpolation.

polyglot.t("hello_name", {name: "Spike"});
=> "Hello, Spike"

Pass a number as the second argument as a shortcut to smart_count:

// same as: polyglot.t("car", {smart_count: 2}); 
polyglot.t("car", 2);
=> "2 cars"

If you like, you can provide a default value in case the phrase is missing. Use the special option key "_" to specify a default.

polyglot.t("i_like_to_write_in_language", {
  _: "I like to write in %{language}.",
  language: "JavaScript"
=> "I like to write in JavaScript."


Related projects

  • i18n-extract: Manage localization with static analysis. (E.g. key usage extraction)