Add missing strings to an existing svelte-i18n setup

An existing svelte-i18n app has its own catalog files, its own init module, its own conventions. The only correct automation is one that joins your world: Polyglot statically resolves the JSON files from your register('en', () => import('./locales/en.json')) calls and merges new keys into those files.

The transformation

<script>
  import { _ } from 'svelte-i18n';
</script>

<h1>Welcome back</h1>
<p>{$_('nav.home')}</p>

<script>
  import { _, t } from 'svelte-i18n';
</script>

<h1>{$t('src.routes.+page.welcome_back')}</h1>
<p>{$_('nav.home')}</p>

Note the import: your file already imported { _ } from svelte-i18n, so Polyglot extended that import with t (its alias) rather than duplicating or breaking it.

Merged into YOUR catalogs

{
  "nav": {
    "home": "Startseite"
  },
  "src": {
    "routes": {
      "+page": {
        "welcome_back": "Welcome back"
      }
    }
  }
}

{
  "nav": {
    "home": "Home"
  },
  "src": {
    "routes": {
      "+page": {
        "welcome_back": "Welcome back"
      }
    }
  }
}

• Key order and formatting are preserved — merges are textual, never a re-serialization that scrambles your diff.
• Every other locale you register gets an English stub for the new key so nothing renders raw until you run polyglot translate.
• If your dictionaries are inline or your loader paths are computed, Polyglot flags the strings with a recipe instead of writing files your app would never load.

Try it

$ curl -fsSL https://getpolyglot.ai/install.sh | bash
$ polyglot init --framework sveltekit
$ polyglot scan
$ polyglot wrap
$ polyglot translate --languages de   # 50 strings free, no signup