Skip to main content

Supported languages

Some voice models also support multiple speakers, which significantly increases the variety of available voices - for example, English LibriTTS alone includes more than 900 different speakers.

This is the complete list of languages supported out of the box, with pre-included models ready to download:

  • English (United States) – 43 voice models, 3 qualities, 1895 speakers (with Kokoro models)
  • English (British) – 19 voice models, 3 qualities, 143 speakers (with Kokoro models)
  • Simplified Chinese (简体中文) – 109 voice models, 2 qualities, 110 speakers (with Kokoro models)
  • Spanish Mexican (Español Mexicano) – 2 voice models, 2 qualities, 2 speakers
  • Spanish European (Español Europeo) – 10 voice models, 2 qualities, 11 speakers (with Kokoro models)
  • Spanish Argentinian (Español Argentino) – 1 voice model, 1 quality, 1 speaker
  • Spanish Colombian (Español Colombiano) – 2 voice models, 1 quality, 2 speakers
  • Korean (한국어) – 1 voice model, 1 quality, 1 speaker
  • Russian (Русский) – 4 voice models, 1 quality, 4 speakers
  • Portuguese Brazil (Português do Brasil) – 9 voice models, 2 qualities, 9 speakers (with Kokoro models)
  • Portuguese Portugal (Português de Portugal) – 3 voice models, 1 quality, 2 speakers
  • Hindi (हिन्दी) – 7 voice models, 1 quality, 7 speakers (with Kokoro models)
  • Urdu (اردو) – 1 voice model, 1 quality, 1 speaker
  • Telugu (తెలుగు) – 3 voice models, 1 quality, 3 speakers
  • Malayalam (മലയാളം) – 2 voice models, 1 quality, 2 speakers
  • German (Deutsch) – 10 voice models, 3 qualities, 254 speakers
  • French (Français) – 8 voice models, 2 qualities, 134 speakers (with Kokoro models)
  • Arabic (العربية) – 4 voice models, 1 quality, 4 speakers
  • Turkish (Türkçe) – 3 voice models, 1 quality, 3 speakers
  • Polish (Polski) – 5 voice models, 3 qualities, 5 speakers
  • Italian (Italiano) – 6 voice models, 2 qualities, 6 speakers (with Kokoro models)
  • Ukrainian (Украї́нська мо́ва) – 5 voice models, 3 qualities, 7 speakers
  • Catalan (Català) – 2 voice models, 2 qualities, 3 speakers
  • Bulgarian (Български) – 1 voice model, 1 quality, 1 speaker
  • Czech (Čeština) – 1 voice model, 2 qualities, 2 speakers
  • Welsh (Cymraeg) – 2 voice models, 1 quality, 8 speakers
  • Danish (Dansk) – 2 voice models, 1 quality, 2 speakers
  • Greek (Ελληνικά) – 1 voice model, 1 quality, 1 speaker
  • Persian (فارسی) – 5 voice models, 1 quality, 5 speakers
  • Finnish (Suomi) – 1 voice model, 2 qualities, 2 speakers
  • Hungarian (Magyar) – 3 voice models, 1 quality, 3 speakers
  • Basque (Euskara) – 4 voice models, 1 quality, 4 speakers
  • Icelandic (Íslenska) – 4 voice models, 1 quality, 4 speakers
  • Indonesian (Bahasa Indonesia) – 1 voice model, 1 quality, 1 speaker
  • Georgian (ქართული ენა) – 1 voice model, 1 quality, 1 speaker
  • Kazakh (Қазақша) – 3 voice models, 2 qualities, 8 speakers
  • Luxembourgish (Lëtzebuergesch) – 1 voice model, 1 quality, 1 speaker
  • Latvian (Latviešu) – 1 voice model, 1 quality, 1 speaker
  • Nepali (नेपाली) – 2 voice models, 2 qualities, 37 speakers
  • Dutch Belgium (Vlaams) – 2 voice models, 2 qualities, 4 speakers
  • Dutch Netherlands (Nederlands) – 8 voice models, 2 qualities, 59 speakers
  • Norwegian (Bokmål / Nynorsk) – 1 voice model, 1 quality, 1 speaker
  • Romanian (Română) – 1 voice model, 1 quality, 1 speaker
  • Slovak (Slovenčina) – 1 voice model, 1 quality, 1 speaker
  • Slovenian (Slovenščina) – 1 voice model, 1 quality, 1 speaker
  • Serbian (Srpski) – 1 voice model, 1 quality, 2 speakers
  • Swedish (Svenska) – 4 voice models, 1 quality, 4 speakers
  • Albanian (Shqip) – 1 voice model, 1 quality, 1 speaker
  • Swahili (Kiswahili) – 1 voice model, 1 quality, 1 speaker
  • Kurdish Kurmanji (Kurdî) – 1 voice model, 1 quality, 2 speakers
  • Vietnamese (Tiếng Việt) – 3 voice models, 2 qualities, 67 speakers

Out of the box, the plugin ships with pre-included models for 51 languages, offering 317 voice models, 75 qualities, and 2834 speakers, among which 151 Kokoro models are available. Additional languages can be added by importing custom voice models (see below).

Adding more voices and languages with custom models​

The list above only covers the voices and languages that ship with pre-included, ready-to-download models. It is not a hard limit on what the plugin can speak.

You can also import any Piper-compatible voice model, whether you download one made by the community or train your own. This lets you add new voices for languages that are already supported, or bring in entirely new languages that aren't shipped out of the box. See Importing custom voice models for the full walkthrough.

Beyond the languages listed above, the plugin can also synthesize speech in the following languages, provided you supply a compatible voice model:

Afrikaans, Albanian, Amharic, Aragonese, Armenian, Assamese, Azerbaijani, Bashkir, Belarusian, Bengali, Bishnupriya Manipuri, Bosnian, Burmese, Chuvash, Croatian, Estonian, Gaelic (Irish & Scottish), Greenlandic, Guarani, Gujarati, Haitian Creole, Hawaiian, Hebrew, Kannada, Konkani, Kyrgyz, Latgalian, Lithuanian, Lule Saami, Macedonian, Malay, Maltese, Māori, Marathi, Nahuatl (Classical), Nogai, Oriya, Oromo, Papiamento, Punjabi, Quechua, Setswana, Shan (Tai Yai), Sindhi, Sinhala, Tamil, Tatar, Thai, Turkmen, Uyghur, and Uzbek.

Since none of these have a pre-trained model bundled with the plugin, you'll need to either train your own or find a community-trained one, for example on Hugging Face, and then import it the same way.