Audio & Speech

This page covers speech generation, transcription, and audio translation endpoints.

Apifox Coverage Notes

• Native OpenAI-style endpoints: TTS, whisper-1, gpt-4o-transcribe, audio translation, audio output, audio input
• Vendor extensions: MiniMax speech synthesis and Doubao voice 2.0
• Native Gemini audio generation and speech flows

Model APIs / Audio

POST Native Gemini Format

POST /v1beta/models/{model}:generateContent

Gemini audio-generation endpoint. You can use models such as gemini-2.5-flash-preview-tts.

Body Parameters

{
  "contents": [
    {
      "parts": [
        {
          "text": "Say cheerfully: Have a wonderful day!"
        }
      ]
    }
  ],
  "generationConfig": {
    "responseModalities": [
      "AUDIO"
    ],
    "speechConfig": {
      "voiceConfig": {
        "prebuiltVoiceConfig": {
          "voiceName": "Kore"
        }
      }
    }
  }
}

Request Parameters

Name	In	Type	Required	Description
model	path	string	Yes	Model Name
key	query	string	Yes	none
Content-Type	header	string	Yes	none
body	body	object	No	none
» contents	body	[object]	Yes	none
»» parts	body	[object]	No	none
»»» text	body	string	No	none

Response Example

200 Response

{
  "id": "string",
  "object": "string",
  "created": 0,
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "string",
        "content": "string"
      },
      "finish_reason": "string"
    }
  ],
  "usage": {
    "prompt_tokens": 0,
    "completion_tokens": 0,
    "total_tokens": 0
  }
}

Responses

Status	Meaning	Description	Model
200	OK	none	Inline

Response Schema

Status 200

Name	Type	Required	Constraints	Chinese Name	Description
» id	string	true	none		none
» object	string	true	none		none
» created	integer	true	none		none
» choices	[object]	true	none		none
»» index	integer	false	none		none
»» message	object	false	none		none
»»» role	string	true	none		none
»»» content	string	true	none		none
»» finish_reason	string	false	none		none
» usage	object	true	none		none
»» prompt_tokens	integer	true	none		none
»» completion_tokens	integer	true	none		none
»» total_tokens	integer	true	none		none

Model APIs / Audio / Native OpenAI Format

POST Doubao Speech 2.0

POST /audio/speech

Different voices support different emotional ranges. For the full list of voice emotions, see: https://www.volcengine.com/docs/6561/1257544 Examples include happy, sad, angry, surprised, and fear... Any parameters outside the OpenAI standard should be passed in metadata using the original vendor format; all vendor-specific fields are supported. Voice-emotion example: adjust the emotion parameter.

Body Parameters

{
    "input": "你是一个好孩子",
    "model": "seed-tts-1.1",
    "response_format": "mp3",
    "speed": 1,
    "voice": "zh_female_shuangkuaisisi_emo_v2_mars_bigtts",
    "metadata": {
        "audio": {
            "emotion_scale": 5,
            "enable_emotion": true,
            "emotion": "happy"
        }
    }
}

Request Parameters

Name	In	Type	Required	Description
Content-Type	header	string	Yes	none
Authorization	header	string	Yes	none
body	body	object	No	none
» model	body	string	Yes	One of the available TTS models: tts-1 or tts-1-hd.
» input	body	string	Yes	Text to synthesize into audio. Maximum length: 4096 characters.
» voice	body	string	Yes	Voice used to generate the audio. Supported voices: alloy, echo, fable, onyx, nova, and shimmer.
» response_format	body	string	No	Audio format to return. Defaults to mp3. Supported formats: mp3, opus, aac, and flac.
» speed	body	number	No	Playback speed for generated audio. Choose a value between 0.25 and 4.0. The default is 1.0.

Response Example

200 Response

{}

Responses

Status	Meaning	Description	Model
200	OK	none	Inline

Response Schema

POST Speech to Text / gpt-4o-transcribe

POST /audio/transcriptions

Official documentation：https://platform.openai.com/docs/api-reference/audio/createTranscription

Body Parameters

file: file:///Users/jun/Downloads/response.wav
model: gpt-4o-transcribe
prompt: eiusmod nulla
response_format: json
temperature: 0
language: ""

Request Parameters

Name	In	Type	Required	Description
Content-Type	header	string	Yes	none
Accept	header	string	Yes	none
Authorization	header	string	No	none
body	body	object	No	none
» file	body	string(binary)	Yes	Audio file to transcribe. Supported formats: mp3, mp4, mpeg, mpga, m4a, wav, or webm.
» model	body	string	Yes	ID of the model to use. Currently only whisper-1 is available.
» prompt	body	string	No	Optional text used to guide the model style or continue a previous audio segment. The prompt should match the language of the audio.
» response_format	body	string	No	none
» temperature	body	number	No	Sampling temperature between 0 and 1. Higher values such as 0.8 make output more random, while lower values such as 0.2 make it more focused and deterministic. If set to 0, the model automatically raises the temperature using log probabilities until a threshold is reached.
» language	body	string	No	Language of the input audio. Providing the language in ISO-639-1 format improves both accuracy and latency.

Response Example

200 Response

{
  "text": "string"
}

Responses

Status	Meaning	Description	Model
200	OK	none	Inline

Response Schema

Status 200

Name	Type	Required	Constraints	Chinese Name	Description
» text	string	true	none		none

POST Audio Translation

POST /audio/translations

Translates audio into English.

Body Parameters

file: string
model: whisper-1
prompt: ""
response_format: verbose_json
temperature: 0

Request Parameters

Name	In	Type	Required	Description
Content-Type	header	string	Yes	none
Accept	header	string	Yes	none
Authorization	header	string	No	none
body	body	object	No	none
» file	body	string(binary)	Yes	Audio file to translate. Supported formats: mp3, mp4, mpeg, mpga, m4a, wav, or webm.
» model	body	string	Yes	ID of the model to use. Currently only whisper-1 is available.
» prompt	body	string	No	Optional text used to guide the model style or continue a previous audio segment. The prompt must be in English.
» response_format	body	string	No	Output format for the transcription. Supported values: json, text, srt, verbose_json, or vtt.
» temperature	body	number	No	Sampling temperature between 0 and 1. Higher values such as 0.8 make output more random, while lower values such as 0.2 make it more focused and deterministic. If set to 0, the model automatically raises the temperature using log probabilities until a threshold is reached.

Response Example

200 Response

{
  "text": "string"
}

Responses

Status	Meaning	Description	Model
200	OK	none	Inline

Response Schema

Status 200

Name	Type	Required	Constraints	Chinese Name	Description
» text	string	true	none		none