Text-to-Speech API

The Text-to-Speech API synthesizes natural Russian speech using Google Cloud Gemini TTS (gemini-3.1-flash-tts-preview, voice Achernar).

Base URL

https://api.koveh.com/tts/

Authentication

Authorization: Bearer <api_key> or X-API-Key: <api_key>

api_users.allowed_models must include tts or text-to-speech.

Endpoints

Health

GET /health — no auth

Presets

GET /presets — list built-in voice presets

Synthesize

POST /synthesize

JSON body:

{
  "text": "Текст для озвучки.",
  "preset": "achernar-ru-zumer",
  "audio_encoding": "LINEAR16",
  "speaking_rate": 1,
  "pitch": 0,
  "response_format": "binary"
}

Returns WAV audio (LINEAR16) or JSON with base64 audio_content when response_format=json.

Default preset prompt: зумерский разговорный голос (Кофемания, Москва), без дикторской подачи.

Example

curl -X POST "https://api.koveh.com/tts/synthesize" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text":"Привет, это тест."}' \
  --output speech.wav