Tavus stock voices in 16 languages vs Sume TTS language mismatch 409

Tavus auto-localized 16 new stock voices (Sep 9, 2026). In Sume, a voice and text language mismatch returns 409 unless you confirm. How to avoid wasted renders.

4 min readSume
All posts

Tavus's changelog entry for September 9, 2026 says it added 16 new stock voices that auto-localize to 16 languages, and that all 42 supported languages are covered. Sume treats language mismatch as a hard check: if the text language does not fit the voice, the TTS call returns 409 tts_voice_language_mismatch, and you override it explicitly with confirm_language_mismatch.

The facts

Tavus facts are from its changelog; the Sume behavior is from its TTS API notes.

Voice and language handling (read 2026-10-02)
FactValue
Tavus new stock voices16, auto-localized to 16 languages
Tavus supported languages42
Sume mismatch response409 tts_voice_language_mismatch
Sume overrideconfirm_language_mismatch
Sume TTS price$0.0475 per 1,000 characters

Why the hard stop helps

A mismatch is usually a bug in a pipeline: Korean text sent to an English voice, or a template that did not switch voice per locale. A 409 stops it before you pay for a render. Override only when the accent is intended.

Related caption behavior is similar: Korean text on a Latin caption style returns 400 caption_hangul_text_latin_style instead of producing a broken video.

A localization checklist

Pair each locale with one voice in config, test the first line with TTS at fractions of a cent, then submit the avatar video. Check the avatar video docs for the caption language field.

Auto-localization is convenient but not the same as a human read of your script, so review audio for names and numbers.

Sources

Related posts

More in Developers

All Developers posts

Written by Sume