Translating HTML and JavaScript using Weblate CDN

Starting with Weblate 4.2 it is possible to export localization to a CDN using JavaScript localization CDN add-on.


This feature is configured on Hosted Weblate. It requires additional configuration on your installation, see LOCALIZE_CDN_URL and LOCALIZE_CDN_PATH.

Upon installation into your component it will push committed translations (see Lazy commits) to the CDN and these can be used in your web pages to localize them.

Creating component

First, you need to create a monolingual component which will hold your strings, see Adding translation projects and components for generic instructions on that.

In case you have existing repository to start with (for example the one containing HTML files), create an empty JSON file in the repository for the source language (see Source language), for example locales/en.json. The content should be {} to indicate an empty object. Once you have that, the repository can be imported into Weblate and you can start with an add-on configuration.


In case you have existing translations, you can place them into the language JSON files and those will be used in Weblate.

For those who do not want to use existing repository (or do not have one), choose Start from scratch when creating component and choose JSON file as a file format (it is okay to choose any monolingual format at this point).

Configuring Weblate CDN add-on

The JavaScript localization CDN add-on provides few configuration options.

Translation threshold

Translations translated above this threshold will be included in the CDN.

CSS selector

Configures which strings from the HTML documents are translatable, see String extraction for Weblate CDN and HTML localization using Weblate CDN.

Language cookie name

Name of cookie which contains user selected language. Used in the JavaScript snippet for HTML localization using Weblate CDN.

Extract strings from HTML files

List of files in the repository or URLs where Weblate will look for translatable strings and offer them for a translation, see String extraction for Weblate CDN.

String extraction for Weblate CDN

The translation strings have to be present in Weblate. You can either manage these manually, use API to create them or list files or URLs using Extract strings from HTML files and Weblate will extract them automatically. The files have to present in the repository or contain remote URLs which will be download and parsed regularly by Weblate.

The default configuration for CSS selector extracts elements with CSS class l10n, for example it would extract two strings from following snippets:

<section class="content">
    <div class="row">
        <div class="wrap">
            <h1 class="section-title min-m l10n">Maintenance in progress</h1>
            <div class="page-desc">
                <p class="l10n">We're sorry, but this site is currently down for maintenance.</p>

In case you don’t want to modify existing code, you can also use * as a selector to process all elements.


Right now, only text of the elements is extracted. This add-on doesn’t support localization of element attributes or elements with children.

HTML localization using Weblate CDN

To localize a HTML document, you need to load the weblate.js script:

<script src="" async></script>

Upon loading, this will automatically find all matching translatable elements (based on CSS selector configuration) and replace their text with a translation.

The user language is detected from the configured cookie and falls back to user preferred languages configured in the browser.

The Language cookie name can be useful for integration with other applications (for example choose django_language when using Django).

JavaScript localization

The individual translations are exposed as bilingual JSON files under the CDN. To fetch one you can use following code:

  .then(response => response.json())
  .then(data => console.log(data));

The actual localization logic needs to be implemented in this case.