---
title: "Preview Tts Voice"
method: GET
path: "/v1/admin/system/tts-preview"
tags: ["admin"]
---

# Preview Tts Voice

`GET /v1/admin/system/tts-preview`

Generate a short audio preview of a TTS voice.

Accepts the new ``provider``/``voice_id`` form, and the legacy ``model``
parameter (interpreted as Deepgram) until the admin UI is updated.

If ``text`` is provided, that phrase is rendered through the chosen
voice — this is how the admin phrase-tester auditions arbitrary text
(punctuation, emphasis markup, etc.) against any voice + language. If
``text`` is absent, the canned ``PREVIEW_TEXTS[lang]`` greeting is used
(preserving the legacy voice-picker preview behavior).

## Query parameters

- `lang` string, required — Language code, e.g. en
- `provider` string, nullable — Provider name (deepgram, elevenlabs, …)
- `voice_id` string, nullable — Provider-specific voice id
- `model` string, nullable — Legacy: full Deepgram model id
- `text` string, nullable — Optional phrase to render. When provided, replaces the canned per-language preview text. Used by the admin phrase-tester.

## Response `200`

Successful Response

- unknown

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.dev/roxels/apis/interviewer.md) · [All operations](https://skmtc.dev/roxels/apis/interviewer/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/roxels/interviewer/revisions/4ef48ed226e1/schema)