List Agent TTS Models

Beta

GET

https://api.speechify.ai/v1/agents/tts-models

GET

/v1/agents/tts-models

curl https://api.speechify.ai/v1/agents/tts-models \
     -H "Authorization: Bearer <token>"

{
  "models": [
    {
      "id": "simba-3.0",
      "display_name": "Simba 3.0",
      "description": "The default for voice agents. Streaming-native, and the only model with a multilingual deployment, so every voice and every language can serve on it.",
      "beta": false,
      "default": true,
      "english_only": false,
      "requires_voice_curation": false
    },
    {
      "id": "simba-3.2",
      "display_name": "Simba 3.2",
      "description": "The richest expressivity of the Simba 3 line. English only, and offered on every Speechify voice, cloned voices included.",
      "beta": false,
      "default": false,
      "english_only": true,
      "requires_voice_curation": false
    },
    {
      "id": "simba-3.5-turbo",
      "display_name": "Simba 3.5 Turbo",
      "description": "Our fastest model - the lowest time-to-first-byte we serve - and the newest English training. English only, and offered on every Speechify voice. Experimental, so evaluate it on a single agent before moving a fleet.",
      "beta": true,
      "default": false,
      "english_only": true,
      "requires_voice_curation": false
    }
  ]
}

List the Simba text-to-speech models a voice agent can be pinned to with tts.model.

The response is the whole catalog and is the same for every workspace - which model serves an agent is a routing decision, not an entitlement, so nothing here is plan-gated. Whether a particular agent may select an entry depends on two facts this response publishes rather than resolving for you, because both change without this list changing:

english_only - the model has no multilingual deployment, so an agent that declares additional_languages (or whose language is not English) cannot select it.
requires_voice_curation - deprecated and always false. No model asks a voice for a registration: every training conditions on the voice’s own prompt audio, so every model is selectable on every voice we speak ourselves, cloned voices included.

Drive a model picker from this and the agent’s languages and it can only offer combinations POST/PATCH /v1/agents will accept. Returns the full set in one response: bounded by the platform’s curated model catalog, so it is intentionally not paginated.

List the Simba text-to-speech models a voice agent can be pinned to with `tts.model`. The response is the whole catalog and is the same for every workspace - which model serves an agent is a routing decision, not an entitlement, so nothing here is plan-gated. Whether a *particular* agent may select an entry depends on two facts this response publishes rather than resolving for you, because both change without this list changing: - `english_only` - the model has no multilingual deployment, so an agent that declares `additional_languages` (or whose `language` is not English) cannot select it. - `requires_voice_curation` - deprecated and always false. No model asks a voice for a registration: every training conditions on the voice's own prompt audio, so every model is selectable on every voice we speak ourselves, cloned voices included. Drive a model picker from this and the agent's languages and it can only offer combinations POST/PATCH /v1/agents will accept. Returns the full set in one response: bounded by the platform's curated model catalog, so it is intentionally not paginated.

Authentication

AuthorizationBearer

Enter your API key with the Bearer prefix, e.g. ‘Bearer sk_…’.

Response headers

Speechify-Request-IdstringOptional

Unique identifier for this request, present on every response (2xx and non-2xx alike). If the caller sends a Speechify-Request-Id request header the server echoes it back (sanitized and length-capped) so one logical request can be traced end-to-end; otherwise the server generates a fresh value. Log it on every response and quote it in support requests

it is the stable handle that ties your observation to Speechify’s server-side logs, and it matches the request_id field in the error envelope.

The legacy alias X-Request-ID carries the same value and is still accepted on requests, until 2027-07-24. Prefer the un-prefixed name (RFC 6648).

Unique identifier for this request, present on every response (2xx and non-2xx alike). If the caller sends a `Speechify-Request-Id` request header the server echoes it back (sanitized and length-capped) so one logical request can be traced end-to-end; otherwise the server generates a fresh value. Log it on every response and quote it in support requests - it is the stable handle that ties your observation to Speechify's server-side logs, and it matches the `request_id` field in the error envelope. The legacy alias `X-Request-ID` carries the same value and is still accepted on requests, until 2027-07-24. Prefer the un-prefixed name (RFC 6648).

Response

The TTS models selectable for voice agents.

modelslist of objects

Errors

401

Unauthorized Error

Authentication

Headers

Response headers

Response

Errors