Text-to-Speech

Use the MiniMax text-to-speech (TTS) capability with both synchronous and asynchronous calling styles. All requests are forwarded to MiniMax by GregAPI — no self-signing required, just pass the GregAPI API Key in the header.

export BASE_URL="https://api.gregapi.com"
export TOKEN="oh-xxxxxxxxxxxxxxxx"

Standard headers:

Authorization: Bearer <TOKEN>
Content-Type: application/json

Endpoints

Operation	Method	Endpoint
Synchronous synthesis	`POST`	`/minimaxi/v1/t2a_v2`
Asynchronous synthesis	`POST`	`/minimaxi/v1/t2a_async_v2`
Asynchronous query	`GET`	`/minimaxi/v1/query/t2a_async_query_v2`

Supported Models

Model	Notes
`speech-2.8-hd`	Recommended default, high-fidelity audio
`speech-2.8-turbo`	Low latency, fast synthesis
`speech-2.6-hd`	Previous-generation high-fidelity model

Request Parameters

Parameter	Type	Required	Description
`model`	string	Yes	Model name, e.g. `speech-2.8-hd`
`text`	string	Yes	Text to synthesize
`voice_id`	string	Yes	Voice ID, or a built-in platform alias
`output_format`	string	No	`hex` (default) or `url`; non-streaming only, `url` validity follows the official return
`emotion`	string	No	Emotion, see the enum below
`sound_effects`	string	No	Sound effect, see the enum below

Emotion enum

happy, sad, angry, fearful, disgusted, surprised, calm, fluent, whisper Support depends on the model and voice combination; not all voice_id values support every emotion.

Sound effects enum

spacious_echo, auditorium_echo, lofi_telephone, robotic

Voice aliases

For OpenAI-style consistency, the platform maps common aliases to MiniMax voices:

Alias	MiniMax voice
`alloy`	`female-chengshu` (mature female)
`echo`	`male-qn-qingse` (youth male - clear)
`fable`	`male-qn-jingying` (youth male - elite)
`onyx`	`presenter_male`
`nova`	`presenter_female`
`shimmer`	`audiobook_female_1`

When you use one of these aliases in voice_id, the platform converts it to the actual MiniMax voice, and fills in the default emotion automatically when the voice has one and no emotion is specified.

Return format (audio_mode)

The audio return mode can be set via channel custom parameters:

audio_mode = json (default): the response body is JSON, with data.audio as hex or URL (determined by output_format).
audio_mode = hex: returns a raw audio stream when output_format=hex; still returns JSON when output_format=url.

Synchronous example

curl -X POST "https://api.gregapi.com/minimaxi/v1/t2a_v2" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "speech-2.8-hd",
    "text": "Welcome to GregAPI",
    "output_format": "hex"
  }' \
  --output minimax-tts.mp3

With output_format=hex, the response is a raw hex audio stream that can be written directly to an .mp3 file. With output_format=url, the response is JSON; read the audio URL from data.audio.

Asynchronous example

curl -X POST "https://api.gregapi.com/minimaxi/v1/t2a_async_v2" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "speech-2.8-hd",
    "text": "hello"
  }'

Query the result:

curl "https://api.gregapi.com/minimaxi/v1/query/t2a_async_query_v2?task_id=$TASK_ID" \
  -H "Authorization: Bearer $TOKEN"

Async synthesis preserves your voice_id and parameters in the JSON payload; if you used an alias, the platform converts it and fills in the emotion before the upstream request.

Billing

Billed by the audio successfully synthesized; see the MiniMax speech entries under “Model Pricing” in the console for exact prices.
Actual amounts follow the upstream consumption logs.

Large Language Models (LLM)

Image Models

Video Models

Speech Models

Endpoints

Supported Models

Request Parameters

Emotion enum

Sound effects enum

Voice aliases

Return format (audio_mode)

Synchronous example

Asynchronous example

Billing

​Endpoints

​Supported Models

​Request Parameters

​Emotion enum

​Sound effects enum

​Voice aliases

​Return format (audio_mode)

​Synchronous example

​Asynchronous example

​Billing

Endpoints

Supported Models

Request Parameters

Emotion enum

Sound effects enum

Voice aliases

Return format (audio_mode)

Synchronous example

Asynchronous example

Billing