OpenAI Audio Transcriptions API
POSThttps://api.acedata.cloud/v1/audio/transcriptions
POSThttps://api.acedata.cloud/openai/audio/transcriptions

Agent integrations

Request Headers

authorizationstring
Bearer token

Request Body

filestringRequired parameter
Audio files to be transcribed, up to 25 MB. Supports flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm formats.
    modelstring
    Transcription model. whisper-1 (default) supports subtitle output and word-level timestamps; gpt-transcribe has more accurate recognition, lower cost, supports languages[] and keywords[], but only returns json/text.
    Please select
    promptstring
    Optional prompt words, used to guide the model's writing style, or provide proper nouns and terminology to enhance recognition accuracy.
    streamboolean
    Whether to return in a streaming manner. whisper-1 does not support streaming; it will be ignored after being passed in (consistent with OpenAI's official behavior) and will still return the complete result.
    Please select
    languagestring
    Audio language, use ISO-639-1 codes (e.g., zh, en). Filling in can improve accuracy and speed; leaving it blank will result in automatic recognition.
    keywords[]array
    List of proprietary terms/terminology prompts, supported only by gpt-transcribe, which can significantly improve the recognition accuracy of brand names, personal names, etc.
    languages[]array
    Candidate language list (ISO-639-1), only supported by gpt-transcribe. Filling in can improve accuracy; it is mutually exclusive with language, do not provide both at the same time.
    temperaturenumber
    Sampling temperature, range 0–1, default 0. The higher the value, the more random the result.
    response_formatstring
    Return format. whisper-1 supports json (default), text, srt, verbose_json, vtt, among which srt and vtt directly return subtitle text; gpt-transcribe only supports json and text.
    Please select
    timestamp_granularities[]array
    Timestamp granularity, optional word (word level) or segment (sentence level), must be used in conjunction with response_format=verbose_json.

    Response

    Integration guide

    Shell

    Python

    JavaScript

    Java

    Go

    PHP

    Kind reminder: For streaming requests, the above code may not be fully applicable. Please refer to the integration documentation for changes.

    OpenAI generation
    Allow Use General Balance

    When 'Allow General Balance' is enabled, the general balance is used automatically if an app's balance is insufficient.