How do I know a new TTS model is available on Sume?

How can you tell whether a newly launched TTS model works on Sume? Call GET /v1/tts-router/models. If the id is not in that list, a request with it returns 400 model_not_found. Blog posts and launch threads, including this one, go stale; the endpoint does not.

In the last few weeks Microsoft released MAI-Voice-2.1, ElevenLabs released Eleven v4 and Mistral released Voxtral TTS (all read 2026-10-04). The Sume router catalog documented in the API contract at the time of reading lists Sonic versions only, so none of those three is selectable there.

Read the catalog

The response is a list of model ids with the capabilities each supports. Fetch it at startup or on a schedule instead of copying ids into code.

import os, requests

r = requests.get("https://api.sume.com/v1/tts-router/models",
                 headers={"Authorization": "Bearer " + os.environ["SUME_API_KEY"]})
r.raise_for_status()
j = r.json()
print(j.get("data", j))

What the contract says about ids

These rules come from the OpenAPI contract behind the API reference.

Keep your integration calm

Pin concrete ids in production, and use the catalog to alert when a new one appears so a person can audition it. Treat an alias such as sonic-latest as a convenience for experiments only, because its meaning moves.

If you need a model that is not listed, the honest answer is that Sume does not offer it yet; use that vendor's own API for that voice, and keep Sume for the parts the voice feeds, such as captions, timelines and avatars.