Non-English Language Support

Non-English Language Support

Overview

When speaking non-English languages, you must specify both voice and tts_provider for proper pronunciation. Default OpenAI voices speak non-English with an American accent.

Recommended Kokoro Voices

Spanish

converse("¿Cómo estás?", voice="ef_dora", tts_provider="kokoro")

French

converse("Bonjour!", voice="ff_siwis", tts_provider="kokoro")

Italian

converse("Ciao!", voice="if_sara", tts_provider="kokoro")

Chinese

converse("你好", voice="zf_xiaobei", tts_provider="kokoro")

Japanese

converse("こんにちは", voice="jf_alpha", tts_provider="kokoro")

Available Kokoro Voices by Language

English

• af_sky - Female
• af_sarah - Female
• am_adam - Male
• am_michael - Male

Spanish

• ef_dora - Female
• em_marco - Male

French

• ff_siwis - Female
• fm_jean - Male

Italian

• if_sara - Female
• im_marco - Male

Chinese (Mandarin)

• zf_xiaobei - Female
• zm_xiaoming - Male

Japanese

• jf_alpha - Female
• jm_beta - Male

German

• gf_lisa - Female
• gm_hans - Male

Portuguese (Brazilian)

• bf_maria - Female
• bm_joao - Male

Russian

• rf_natasha - Female
• rm_dmitri - Male

Korean

• kf_yuna - Female
• km_junho - Male

Hindi

• hf_priya - Female
• hm_raj - Male

Arabic

• af_layla - Female
• am_omar - Male

Turkish

• tf_ayse - Female
• tm_mehmet - Male

Voice Naming Convention

Kokoro voices follow the pattern: {language}{gender}_{name}

• Language code: First letter(s) of language

  • a = English (American)
  • e = Spanish (Español)
  • f = French (Français)
  • i = Italian
  • z = Chinese (中文)
  • j = Japanese
  • g = German
  • b = Brazilian Portuguese
  • r = Russian
  • k = Korean
  • h = Hindi
  • t = Turkish

• Gender:

  • f = Female
  • m = Male

• Name: Given name in that language

OpenAI Voices (English-centric)

OpenAI voices work for any language but maintain American English accent:

• nova - Female
• shimmer - Female
• alloy - Neutral
• echo - Male
• fable - Male
• onyx - Male

Not recommended for non-English - use Kokoro instead.

Important Notes

1. Always specify both voice and tts_provider when using non-English
2. Never use coral voice - it's not supported
3. Kokoro provides native pronunciation for each language
4. STT (speech-to-text) automatically detects language

See Also

• voicemode-parameters - Full parameter reference
• voicemode-quickstart - Basic usage examples