Skip to main content

Voice Support by Language

Speech Mode speaks the translation in every language Whisperr translates into. How much control you have over which voice speaks it depends on the language you are translating into — the spoken language — because different languages are served by different voice catalogues.

For what the settings do, see Choosing the Voice in the Speech Mode overview.

What the Marks Mean​

  • ✅ — supported. Auto picks from 30 natural voices to match the speaker's pitch, timbre, and expressiveness, so each person keeps their own voice; Male and Female pin one fixed voice.
  • ✅* — supported, with a coarser Auto: it follows the speaker's vocal register, so a low voice is read by a male voice and a high one by a female voice, but two men speaking share the same voice.
  • ❌ — not available in this language, so choosing it makes no audible difference.
  • Default — the language's standard voice, which every language has. It reads the translation whenever the setting you picked can't be honoured, so you always hear the translation.
Auto's per-speaker matching runs in the Web App

Matching a voice to each individual speaker needs the acoustic measurements the Web App and Broadcast Mode rooms send with each line. On iPhone, iPad, and Android, Auto uses the standard voice for the target language; Male and Female work everywhere. See Choosing the Voice.

Supported Speech Languages​

Supported speech languageAutoMaleFemaleDefault
Afrikaans✅*✅✅✅
Albanian✅*✅✅✅
Arabic✅✅✅✅
Armenian✅*✅✅✅
Azerbaijani✅*✅✅✅
Basque✅*✅✅✅
Belarusian❌❌❌✅
Bengali✅✅✅✅
Bosnian✅*✅✅✅
Bulgarian✅✅✅✅
Cantonese✅✅✅✅
Catalan✅*✅✅✅
Chinese (Simplified)✅✅✅✅
Chinese (Traditional)✅*✅✅✅
Croatian✅✅✅✅
Czech✅✅✅✅
Danish✅✅✅✅
Dutch✅✅✅✅
English✅✅✅✅
Esperanto❌❌❌✅
Estonian✅✅✅✅
Finnish✅✅✅✅
French✅✅✅✅
Galician✅*✅✅✅
German✅✅✅✅
Greek✅✅✅✅
Gujarati✅✅✅✅
Haitian Creole❌❌❌✅
Hebrew✅✅✅✅
Hindi✅✅✅✅
Hungarian✅✅✅✅
Indonesian✅✅✅✅
Irish✅*✅✅✅
Italian✅✅✅✅
Japanese✅✅✅✅
Kannada✅✅✅✅
Kazakh✅*✅✅✅
Korean✅✅✅✅
Latvian✅✅✅✅
Lithuanian✅✅✅✅
Macedonian✅*✅✅✅
Malay✅*✅✅✅
Malayalam✅✅✅✅
Maltese✅*✅✅✅
Marathi✅✅✅✅
Māori❌❌❌✅
Nepali✅*✅✅✅
Norwegian✅✅✅✅
Persian✅*✅✅✅
Polish✅✅✅✅
Portuguese (Brazil)✅✅✅✅
Portuguese (Portugal)✅*✅✅✅
Punjabi✅✅✅✅
Romanian✅✅✅✅
Russian✅✅✅✅
Serbian✅✅✅✅
Slovak✅✅✅✅
Slovenian✅✅✅✅
Spanish✅✅✅✅
Swahili✅✅✅✅
Swedish✅✅✅✅
Tagalog / Filipino✅*✅✅✅
Tamil✅✅✅✅
Telugu✅✅✅✅
Thai✅✅✅✅
Turkish✅✅✅✅
Ukrainian✅✅✅✅
Urdu✅✅✅✅
Vietnamese✅✅✅✅
Welsh✅*✅✅✅
Wu Chinese✅✅✅✅

* Auto still depends on the speaker's voice — it follows the register their voice sits in, rather than matching them as an individual.

Notes​

  • The language you speak doesn't matter. Voice matching is measured from the sound of your voice, not from what you are saying, so any source language Whisperr transcribes can drive it. Only the language being spoken back decides what the table above says.
  • Regional variants can differ. Brazilian Portuguese gets per-speaker matching; European Portuguese follows the speaker's register instead. Simplified Chinese and Cantonese get per-speaker matching; Traditional Chinese follows the register. Where a specific regional voice doesn't exist, Whisperr uses the closest one for that language.
  • Meeting Bot sessions use the standard voice. A Meeting Bot transcript doesn't carry the per-speaker measurements, so Auto falls back to the language's standard voice there.
  • This list grows. Voice catalogues are checked against the speech providers each time the service starts, so a language can gain support without any app update. Verified September 2026.