MiniMax T2A API

Model Introduction

Exchange Token provides MiniMax text-to-speech through the native endpoint below:

• POST /v1/t2a_v2

Use this endpoint when you want to convert text into audio with MiniMax-compatible request and response formats.

Supported Models

• speech-2.8-hd
• speech-2.8-turbo
• speech-2.6-hd
• speech-2.6-turbo
• speech-02-hd
• speech-02-turbo

Calling Method

API Endpoint

https://api.exchangetoken.ai

Authentication

MiniMax T2A uses the following authentication header:

x-key: YOUR_API_KEY

Content Type

Content-Type: application/json

Request Path

POST /v1/t2a_v2

Core Request Fields

Field	Type	Required	Description
model	string	Yes	TTS model name
text	string	Yes	Input text, recommended within 10000 characters
stream	boolean	No	Whether to return streaming audio chunks
voice_setting	object	Yes	Voice configuration, must include voice_id
audio_setting	object	No	Audio encoding settings
language_boost	string	No	Language hint, for example auto, English, Chinese
pronunciation_dict	object	No	Custom pronunciation mapping
voice_modify	object	No	Pitch, intensity, timbre, and sound effects
subtitle_enable	boolean	No	Whether subtitle output is enabled
subtitle_type	string	No	sentence, word, or word_streaming
output_format	string	No	Non-streaming supports hex and url; streaming returns hex chunks

Non-Streaming Example

curl -X POST "https://api.exchangetoken.ai/v1/t2a_v2" \
  -H "x-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "speech-2.8-hd",
    "text": "Welcome to Exchange Token. This is a MiniMax text to speech demo.",
    "stream": false,
    "language_boost": "auto",
    "output_format": "hex",
    "voice_setting": {
      "voice_id": "English_Graceful_Lady",
      "speed": 1,
      "vol": 1,
      "pitch": 0
    },
    "audio_setting": {
      "sample_rate": 32000,
      "bitrate": 128000,
      "format": "mp3",
      "channel": 1
    }
  }'

Non-Streaming Response Example

{
  "data": {
    "audio": "<hex encoded audio>",
    "status": 2
  },
  "extra_info": {
    "audio_length": 11124,
    "audio_sample_rate": 32000,
    "audio_size": 179926,
    "bitrate": 128000,
    "word_count": 163,
    "invisible_character_ratio": 0,
    "usage_characters": 163,
    "audio_format": "mp3",
    "audio_channel": 1
  },
  "trace_id": "01b8bf9bb7433cc75c18eee6cfa8fe21",
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}

Streaming Example

curl -N -X POST "https://api.exchangetoken.ai/v1/t2a_v2" \
  -H "x-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "speech-2.8-turbo",
    "text": "This request returns audio chunks in streaming mode.",
    "stream": true,
    "output_format": "hex",
    "voice_setting": {
      "voice_id": "English_Insightful_Speaker",
      "speed": 1,
      "vol": 1,
      "pitch": 0
    },
    "audio_setting": {
      "sample_rate": 32000,
      "bitrate": 128000,
      "format": "mp3",
      "channel": 1
    }
  }'

When stream=true, the server returns multiple chunks. Intermediate chunks usually carry data.status = 1, and the final chunk carries data.status = 2.

Streaming Chunk Example

{
  "data": {
    "audio": "<hex encoded audio chunk>",
    "status": 1
  },
  "trace_id": "01b8bf9bb7433cc75c18eee6cfa8fe21",
  "base_resp": {
    "status_code": 0,
    "status_msg": ""
  }
}

Voice and Audio Settings

voice_setting

Field	Type	Required	Description
voice_id	string	Yes	Target voice ID
speed	number	No	Range 0.5 to 2
vol	number	No	Range (0, 10]
pitch	integer	No	Range -12 to 12
emotion	string	No	Emotion style

audio_setting

Field	Type	Required	Description
sample_rate	integer	No	8000, 16000, 22050, 24000, 32000, 44100
bitrate	integer	No	32000, 64000, 128000, 256000
format	string	No	mp3, pcm, flac, wav, pcmu_raw, pcmu_wav, opus
channel	integer	No	1 for mono, 2 for stereo

Response Fields

Field	Description
data.audio	Generated audio content
data.status	1 means generating, 2 means completed
extra_info.usage_characters	Billable character count for this synthesis
trace_id	Request trace ID
base_resp.status_code	Request status code, 0 means success

Error Response Format

{
  "base_resp": {
    "status_code": 1004,
    "status_msg": "authentication failed"
  }
}

Common meanings:

• 0: success
• 1002: rate limit exceeded
• 1004: authentication failed
• 1039: TPM rate limit exceeded
• 1042: invalid characters exceed the limit
• 2013: invalid input parameters

Notes

• For long text, streaming mode is recommended to reduce waiting time for the first audio chunk.
• In non-streaming mode, output_format supports hex and url.
• In streaming mode, audio is returned as chunked hex data.
• trace_id should be kept for troubleshooting.