MiniMax speech synthesis

Pricing

Quick start

1. Replace <API_KEY> in the example with your API key, edit the content and send the request.
2. Save every ID in task_ids and query each one every 3–5 seconds.
3. When status is completed, use the returned file URL to access your result.

Endpoint

POST https://zcbservice.aizfw.cn/kyyReactApiServer/v2/model-center/tasks

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

First-use fee

Character billing

Request parameters

modelstringrequired

Use this model name.

Values and limits: minimax-tts

textstringrequired

The text to read aloud.

Values and limits: At least 1 characters; Up to 9999 characters; Pattern: \S

voice_settingobjectrequired

Voice and reading settings. Start with voice_id.

voice_setting.volnumber

Volume. Higher values are louder.

Values and limits: At most 10; Greater than 0; Documented default (not inserted by the platform): 1

voice_setting.pitchinteger

Pitch. Zero keeps the original pitch.

Values and limits: Range: -12–12; Documented default (not inserted by the platform): 0

voice_setting.speednumber

Reading speed. Higher values are faster.

Values and limits: Range: 0.5–2; Documented default (not inserted by the platform): 1

voice_setting.emotionstring

Speaking emotion. Omit for natural delivery.

Values and limits: Allowed values: "happy"; "sad"; "angry"; "fearful"; "disgusted"; "surprised"; "calm"; "fluent"; "whisper"

voice_setting.voice_idstring

Required when the object is provided

Use an available system voice ID or the voice_id returned by voice cloning. Cloned voices belong to the account that created them.

Values and limits: At least 1 characters; Pattern: \S

voice_setting.latex_readboolean

控制是否朗读 latex 公式，默认为 false 需注意: - 仅支持中文，开启该参数后，language_boost 参数会被设置为 Chinese - 请求中的公式需要在公式的首尾加上 $$ - 请求中公式若有 "\"，需转义成 "\\". 示例：一元二次方程根的基本公式 应表示为 $$x = \\frac{-b \\pm \\sqrt{b^2 - 4ac}}{2a}$$

Values and limits: Documented default (not inserted by the platform): false

voice_setting.text_normalizationboolean

废弃参数，请使用请求体一级参数 text_normalization_mode：本参数为 true 等价于 text_normalization_mode=basic；为 false 或不传等价于不传 text_normalization_mode。两者同时传入时以 text_normalization_mode 为准 是否启用中文、英语文本规范化，开启后可提升数字阅读场景的性能，但会略微增加延迟，默认值为 false

Values and limits: Documented default (not inserted by the platform): false

language_booststring

Reading language; auto detects it automatically.

Values and limits: Allowed values: "Chinese"; "Chinese,Yue"; "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"; "auto"

upstream_modelstringrequired

Speech model version. Use one of the allowed values below.

Values and limits: Allowed values: "speech-2.8-hd"; "speech-2.8-turbo"; "speech-2.6-hd"; "speech-2.6-turbo"; "speech-02-hd"; "speech-02-turbo"; "speech-01-hd"; "speech-01-turbo"

text_normalization_modestring

文本归一化模式。basic 使用快速的规则归一化；quality 采用 LLM + 规则的混合方案，归一化质量更高，但会增加一定延迟。详见文本归一化。 注意： - quality 目前仅对非流式请求（stream=false）生效，且仅支持 speech-2.8-hd、speech-2.8-turbo、speech-2.6-hd、speech-2.6-turbo；流式请求或其他模型传入 quality 时按 basic 处理，不会报错 - 传入本参数后，voice_setting.text_normalization 不再生效；不传时沿用 voice_setting.text_normalization 的设置（默认关闭）

Values and limits: Allowed values: "basic"; "quality"

Request examples

{
  "model": "minimax-tts",
  "text": "你好，这是同步语音合成示例。",
  "voice_setting": {
    "voice_id": "male-qn-qingse"
  },
  "upstream_model": "speech-2.8-hd"
}

Submission response

Query results

GET https://zcbservice.aizfw.cn/kyyReactApiServer/v2/model-center/tasks/{task_id}

Task status

Completed response example

{
  "id": "mcp_00000000000000000000000000000001",
  "model": "minimax-tts",
  "object": "audio",
  "created": 1792022400,
  "status": "completed",
  "amount": 0.25,
  "billing_status": "deducted",
  "refund_status": "none",
  "usage_characters": 0,
  "audio_url": "https://storage.example.com/audio.bin",
  "error": null
}

Response fields

Result fields

Billing and troubleshooting

• Submission timeout: check whether the original request succeeded before submitting again to avoid duplicate generation and charges. If you have a task ID, query that task.
• Temporary query error: retry the query with the same task ID; do not create a new generation.
• delivery_exception: the generation charge is retained. Contact support to restore file access.