Cartesia Sonic 3 languages: 44 in Sonic 3.6, how to set one

Cartesia's Sonic 3.6 text-to-speech model speaks 44 languages. Cartesia's docs list them by base code, from en, fr and ko to two new ones, Odia (or) and Urdu (ur). You choose one with the language or locale field, never both. Sume's TTS request has a language field only.

Cartesia's facts are from its Sonic 3.6 docs and launch post, read 2026-09-29. Sume's are from the TTS request schema in the Sume API reference.

Which 44 languages does Sonic 3.6 speak?

Cartesia's docs table lists these base language codes for the stable snapshot dated August 27, 2026: en, de, es, fr, ja, pt, zh, hi, ko, it, nl, pl, ru, sv, tr, tl, bg, ro, ar, cs, el, fi, hr, ms, sk, da, ta, uk, hu, no, vi, bn, th, he, ka, id, te, gu, kn, ml, mr, pa, or, ur.

Odia and Urdu are new in Sonic 3.6. Cartesia's launch post also says Hindi and English can be mixed in one generation (Hinglish), with the transcript written in Devanagari, Latin script, or a mix.

How do the language and locale fields work?

Per Cartesia's docs, language and locale accept a base code such as en or a regional code such as en-GB. Set one, never both. By default text normalization follows that value, so dates, times and numbers are read the local way; a separate normalization setting lets you use different conventions.

How do I set the language on Sume?

Sume's TTS request has one field for this, language: a BCP-47 or ISO-639 code of 2 to 16 characters, such as ko, ja or en. The schema says to set it for every non-English transcript: omitted, it defaults to English at the provider, and Sume infers ko or ja only as a fallback from a Hangul or kana-only transcript.

The request schema lists no locale or normalization field and does not allow extra fields, so send only language. The Sume docs do not list which language codes each router model accepts; check the audio for any language you rely on.

curl -X POST https://api.sume.com/v1/tts-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: sonic-36-es-001" \
  -d '{
    "model": "sonic-3.6",
    "transcript": "Su pedido sale mañana.",
    "avatar_handle": "acme",
    "language": "es"
  }'

What happens if the voice speaks another language?

The TTS schema has a confirm_language_mismatch boolean: set it to true only after the user confirms a voice-language mismatch warning. It does not change the voice or the language. For fixing wrong words rather than wrong languages, see Text to speech pronunciation.