API docs / Audio models

speech-2.8-turbo API

Convert text into natural, expressive speech with support for over 40 languages and low-latency generation under 250 milliseconds. Features 17 built-in voices including authority figures, friendly voices, energetic options, and character voices, plus support for custom voice cloning. Provides precise control over speech characteristics including speed (0.5x to 2x), pitch adjustment (±12 semitones), volume levels (0-10), and emotional expression with options for happy, sad, angry, fearful, disgusted, surprised, calm, fluent, and neutral delivery. Recognizes over 20 natural interjections like laughs, chuckles, sighs, coughs, gasps, humming, whistles, crying, breath, pant, and sneezes for lifelike speech. Supports pause insertion using markers, custom pronunciations for brand names and specialized terms, and multiple audio formats including MP3, WAV, FLAC, and PCM with configurable sample rates and bitrates. Enables subtitle generation with sentence timestamps and includes language-specific optimizations for improved pronunciation accuracy.

Model id
minimax/speech-2.8-turbo
Provider
minimax
Output
Audio
Price
From 1 credits per run, the same as on voor.ai. Options such as length or resolution can change it.

Request

POST https://voor.ai/v1/generations with "model": "minimax/speech-2.8-turbo". Poll GET /v1/tasks/{id} or receive the signed webhook. See the overview for keys, idempotency, webhooks and errors.

curl https://voor.ai/v1/generations \
  -H "Authorization: Bearer $VOOR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
  "model": "minimax/speech-2.8-turbo",
  "input": {
    "prompt": "A red fox standing in fresh snow at golden hour, soft light"
  },
  "callback_url": "https://example.com/webhooks/voor"
}'

# Then poll the task (or wait for the webhook):
curl https://voor.ai/v1/tasks/TASK_ID \
  -H "Authorization: Bearer $VOOR_API_KEY"

Input parameters

NameTypeDetails
promptrequiredstring

Text to narrate (max 10,000 characters). Use markers like <#0.5#> to insert pauses in seconds.

Up to 10000 characters

voicestring

Voice to synthesize. Pick any MiniMax system voice (e.g. See the full list of voices in the README.

Default: English_Wiselady

output_formatstring

Default: mp3

One of: mp3, wav, flac, pcm

bitratenumber

Default: 128000

One of: 32000, 64000, 128000, 256000

channelstring

Default: mono

One of: mono, stereo

emotionstring

Default: auto

One of: auto, happy, sad, angry, fearful, disgusted, surprised, calm, fluent, neutral

english_normalizationboolean

Improve number/date reading for English text (adds a small amount of latency).

Default: false

language_booststring

Default: None

One of: None, Automatic, Chinese, Chinese,Yue, Cantonese, English, Arabic, Russian, Spanish, French, Portuguese, German, Turkish, Dutch, Ukrainian, Vietnamese, Indonesian, Japanese, Italian, Korean, Thai, Polish, Romanian, Greek, Czech, Finnish, Hindi, Bulgarian, Danish, Hebrew, Malay, Persian, Slovak, Swedish, Croatian, Filipino, Hungarian, Norwegian, Slovenian, Catalan, Nynorsk, Tamil, Afrikaans

pitchnumber

Semitone offset applied to the voice (−12 to +12).

Default: 0

Min -12 · Max 12

sample_ratenumber

Default: 32000

One of: 8000, 16000, 22050, 24000, 32000, 44100

speednumber

Speech speed multiplier (0.5–2.0). Lower is slower, higher is faster.

Default: 1

Min 0.5 · Max 2

subtitle_enableboolean

Return MiniMax subtitle metadata with sentence timestamps (non-streaming only).

Default: false

volumenumber

Relative loudness. 1.0 is default MiniMax gain. Range 0–10.

Default: 1

Min 0 · Max 10