Skip to main content

Supported languages

Tavus supports 43 languages for spoken interaction, powered by three integrated text-to-speech (TTS) engines: Cartesia, ElevenLabs, and Azure. If a selected language is not supported by our default TTS engine (Cartesia), your CVI will automatically switch to ElevenLabs to kick off the conversation.
Language availability also depends on your selected STT model. Some models support a subset of these languages. See the STT layer configuration for per-model language breakdowns.
  • English (en)
  • French (fr)
  • German (de)
  • Spanish (es)
  • Portuguese (pt)
  • Chinese (zh)
  • Japanese (ja)
  • Hindi (hi)
  • Italian (it)
  • Korean (ko)
  • Dutch (nl)
  • Polish (pl)
  • Russian (ru)
  • Swedish (sv)
  • Turkish (tr)
  • Tagalog (tl)
  • Bulgarian (bg)
  • Romanian (ro)
  • Arabic (ar)
  • Czech (cs)
  • Greek (el)
  • Finnish (fi)
  • Croatian (hr)
  • Malay (ms)
  • Slovak (sk)
  • Danish (da)
  • Tamil (ta)
  • Ukrainian (uk)
  • Hungarian (hu)
  • Norwegian (no)
  • Vietnamese (vi)
  • Bengali (bn)
  • Thai (th)
  • Hebrew (he)
  • Georgian (ka)
  • Indonesian (id)
  • Telugu (te)
  • Gujarati (gu)
  • Kannada (kn)
  • Malayalam (ml)
  • Marathi (mr)
  • Punjabi (pa)
  • Swahili (sw)
For a full list of supported languages for each TTS engine, please click on the following links:

Cartesia (default)

ElevenLabs

Azure

By default, Tavus uses the tavus-auto TTS engine, which supports all 42 languages. It chooses the best TTS engine (Cartesia, ElevenLabs, or Azure) for each conversation’s language.

Setting the Conversation Language

To specify a language, use the properties.language parameter in the Create Conversation. You must use the full language name, not a language code.
cURL
Language names must match exactly with those supported by the selected TTS engine.

Smart Language Detection

To automatically detect the participant’s spoken language throughout the conversation, set language to multilingual when creating the conversation:
cURL
This enables the STT (speech-to-text) engine to automatically switch languages, dynamically adjusting the pipeline to transcribe and respond in the detected language throughout the conversation.
For the highest accuracy, we recommend setting a specific language rather than using multilingual. Smart Language Detection works best as a fallback when the participant’s language is unknown ahead of time.