Skip to main content
POST
/
api
/
v2
/
generate
Text-to-Speech
curl --request POST \
  --url https://api.example.com/api/v2/generate \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "text": "<string>",
  "voice_id": "<string>",
  "speed": 123,
  "format": "<string>",
  "audio_url": "<string>",
  "language": "<string>",
  "diarize": true
}
'

​
Request

​
Authorization
string
required
Authorization: Bearer nb_YOUR_API_KEY
​
model
string
required
TTS model slug:
  • minimax-tts — Free, Chinese/English, very natural
  • openai-tts — OpenAI standard voices
  • openai-tts-hd — OpenAI HD quality voices
  • gpt-4o-mini-tts — GPT-4o Mini TTS
  • elevenlabs-flash — ElevenLabs fast (low latency)
  • elevenlabs-v2 — ElevenLabs Multilingual v2 (highest quality)
​
text
string
required
The text to synthesize. Maximum length depends on model (typically 5,000 characters).
​
voice_id
string
Voice identifier. Available voices depend on the model. See examples below.
​
speed
number
default:"1.0"
Speaking speed multiplier. Range: 0.5 – 2.0.
​
format
string
default:"mp3"
Output audio format: mp3, wav, ogg

​
Available Voices

​
OpenAI TTS

alloy, echo, fable, onyx, nova, shimmer

​
ElevenLabs

ElevenLabs supports hundreds of voices. Use common ones like rachel, adam, bella, josh or pass any ElevenLabs voice ID directly.

​
MiniMax TTS

male-qn-qingse, male-qn-jingying, female-shaonv, female-yujie (and more)

​
Request Example

curl -X POST https://elumenta.ru/api/v2/generate \
  -H "Authorization: Bearer nb_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "elevenlabs-v2",
    "text": "Welcome to Elumenta. Your AI-powered creative studio.",
    "voice_id": "rachel",
    "speed": 1.0,
    "format": "mp3"
  }'

​
Response

{
  "id": "gen_01j9x2tts001",
  "status": "completed",
  "model": "elevenlabs-v2",
  "url": "https://storage.elumenta.ru/generations/gen_01j9x2tts001.mp3",
  "duration_seconds": 3.4,
  "characters": 54,
  "tokens_used": 5,
  "balance_remaining": 284,
  "created_at": "2026-03-08T12:10:00Z"
}

​
Speech-to-Text

Transcribe audio files to text using Whisper, GPT-4o Transcribe, or ElevenLabs Scribe.

​
Request

​
model
string
required
STT model:
  • whisper — Fast, multilingual, free
  • gpt-4o-transcribe — Highest accuracy
  • elevenlabs-scribe — Best for podcasts and meetings (diarization support)
​
audio_url
string
required
URL to the audio file (MP3, WAV, OGG, M4A, FLAC). Max 25MB.
​
language
string
ISO 639-1 language code (e.g. en, ru, de). If omitted, the model auto-detects.
​
diarize
boolean
default:"false"
Speaker diarization (who said what). Only available with elevenlabs-scribe.

​
Request Example

curl -X POST https://elumenta.ru/api/v2/generate \
  -H "Authorization: Bearer nb_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "whisper",
    "audio_url": "https://example.com/interview.mp3",
    "language": "en"
  }'

​
Response

{
  "id": "gen_01j9x2stt001",
  "status": "completed",
  "model": "whisper",
  "text": "Hello and welcome to today's episode...",
  "language": "en",
  "duration_seconds": 142.5,
  "tokens_used": 0,
  "created_at": "2026-03-08T12:15:00Z"
}