Gemini TTS regional voices vs Sume's language tag

Sume's TTS language field accepts a BCP-47 or ISO-639 code of 2 to 16 characters, so a tag such as es-MX or fr-CA fits the field. The reference does not say that a regional tag switches the accent or dialect, so treat it as a request and check the audio.

Google's regional claims are from its Gemini 3.8 TTS post. Sume's are from the API reference, read 2026-09-30.

What does Google say about regional varieties?

The post says Gemini 3.8 TTS supports more than 100 languages and dialects and names Mexican Spanish, Quebec French and Scots English as regional varieties. It also lists dialect shifts among the things you can direct line by line.

What does the `language` field accept on Sume?

The schema describes it as the language the voice speaks the transcript in, with ko, ja and en as examples. It is a string with a minimum length of 2 and a maximum of 16.

Set it for every non-English transcript. If you omit it, the reference says it defaults to English at the provider; Sume infers ko or ja only as a fallback from a Hangul-only or kana-only transcript, and the reference says never to translate a non-English request into English.

Will es-MX sound different from es?

The reference does not document a regional accent guarantee for any tag, and it does not list which codes each router model accepts. The reliable test is to render the same sentence with es and es-MX, listen, and keep the one you want.

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: es-mx-test-001" \
  -d '{
    "model": "sonic-3.6",
    "transcript": "Su pedido sale mañana.",
    "avatar_handle": "acme",
    "language": "es-MX"
  }'

Where do I read more about languages on Sume?

The Sonic side of the question, including which languages Cartesia lists, is in Cartesia Sonic 3 languages and the language field. For the Gemini engine question, see which engine Sume routes to.