logoDocs
    Sign UpConsole
    Cloud Real-time ASR
    Product Introduction
    Overview
    Pricing
    Release Notes
    Quick Start
    Quick Integration
    Configure ASR
    Best Practices
    1v1 Real-time Translation Subtitles
    Basic Features
    Display Subtitles
    Enable Translation
    API Reference
    Accessing Methods
    Debug Server API Online
    Return Codes
    POST
    Start Cloud Real-Time ASR
    POST
    Stop Cloud Real-Time ASR
    POST
    Add ASR Stream
    POST
    Delete ASR Stream
    Callbacks
    Receiving Callbacks
    Exception Event Error Codes
    On this page

    StartRealtimeASRTask

    2026-09-18
    POST

    https://cloud-realtime-asr-api.zegotech.cn/

    Start cloud real-time speech recognition task and add the speech recognition instance to the RTC conversation.

    Request

    Query Parameters

      Action stringrequired

      Possible values: [StartRealtimeASRTask]

      Interface prototype parameters

      https://cloud-realtime-asr-api.zegotech.cn?Action=StartRealtimeASRTask

      AppId uint32required

      💡Public parameter. Application ID, assigned by ZEGOCLOUD. Get it from the ZEGOCLOUD Admin Console.

      SignatureNonce stringrequired

      Common parameter. 16-bit hexadecimal random string (hex encoding of 8 bytes of random number). For generation algorithm, refer to Signature Example.

      Timestamp int64required

      Common parameter. Current Unix timestamp, in seconds. For generation algorithm, refer to Signature Example, maximum allowed deviation of 10 minutes.

      SignatureVersion stringrequired

      Possible values: [2.0]

      Default value: 2.0

      💡Public parameter. Signature version number.

      Signature stringrequired

      Common parameter. Signature, used to verify the legitimacy of the request. Please refer to Signature Mechanism to generate.

    Body

    required
      RoomId stringrequired

      RTC room ID

      RecognitionRange integer

      Possible values: [0, 1]

      Recognition range. 0: all streams in the room, 1: specified StreamList.

      ASR object
      ASR parameters
      Vendor string

      Possible values: [Tencent, AliyunParaformer, AliyunFunASR, AliyunQwenAudioASR, VolcSeedASR, Microsoft]

      Default value: Tencent

      ASR vendor. Refer to Configure ASR parameter description. If you need Alibaba, Microsoft recognition vendors, please contact ZEGOCLOUD business support to enable.

      Params object

      Vendor parameters. For specific usage, refer to the Params parameter description in Configure ASR.

      VADSilenceSegmentation number

      Possible values: >= 200 and <= 2000

      Default value: 500

      Used to set how many milliseconds after the user stops speaking, two sentences will no longer be considered as one. Unit is ms, range [200, 2000], default is 500. For detailed explanation, refer to Configure ASR - Speech Segmentation Control.

      HotWord stringdeprecated

      This parameter is deprecated. Please set it through Params vendor parameters.

      StreamList object[]
      List of streams to be recognized, takes effect when RecognitionRange parameter is 1. When RecognitionRange parameter is 1, this parameter cannot be empty.
    • Array[
      • StreamId stringrequired

      Stream ID to be recognized.

      ASR object
      ASR parameters
      Vendor string

      Possible values: [Tencent, AliyunParaformer, AliyunFunASR, AliyunQwenAudioASR, VolcSeedASR, Microsoft]

      Default value: Tencent

      ASR vendor. Refer to Configure ASR parameter description. If you need Alibaba, Microsoft recognition vendors, please contact ZEGOCLOUD business support to enable.

      Params object

      Vendor parameters. For specific usage, refer to the Params parameter description in Configure ASR.

      VADSilenceSegmentation number

      Possible values: >= 200 and <= 2000

      Default value: 500

      Used to set how many milliseconds after the user stops speaking, two sentences will no longer be considered as one. Unit is ms, range [200, 2000], default is 500. For detailed explanation, refer to Configure ASR - Speech Segmentation Control.

      HotWord stringdeprecated

      This parameter is deprecated. Please set it through Params vendor parameters.

      EnableTranslation boolean

      Whether to enable translation.

      Translation object
      Translation LLM configuration item
      Vendor stringrequired

      Possible values: [DoubaoSeedTranslation, QwenMT]

      LLM vendor to call.

      SourceLanguage stringrequired

      Source language

      TargetLanguage stringrequired

      Target language

      LLM objectrequired
      LLM parameters for translation.
      Url stringrequired

      LLM interface address to call. Different for different Vendors:

      • When Vendor is DoubaoSeedTranslation: https://ark.cn-beijing.volces.com/api/v3/responses
      • When Vendor is QwenMT: https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions
      ApiKey stringrequired

      The parameter used for authentication by the LLM service provider.

      Model stringrequired

      The LLM model name to call, required. Different vendors support different models.

      Params object

      LLM or application platform request parameters, optional.

      • For OpenAI protocol, you can configure parameters such as max_tokens, temperature, stream, etc.
      • For other application platforms, you can fill in the parameters supported by the platform.
    • ]
      • AdvancedConfig object
      Advanced configuration parameters.
      MaxIdleTime integer

      Possible values: >= 30 and <= 1800

      Default value: 120

      If no one is publishing streams in the RTC room for more than MaxIdleTime, the background automatically stops the recognition task, in seconds.

      EnableTranslation boolean

      Whether to enable translation. Disabled by default.

      Translation object
      Translation LLM configuration item
      Vendor stringrequired

      Possible values: [DoubaoSeedTranslation, QwenMT]

      LLM vendor to call.

      SourceLanguage stringrequired

      Source language

      TargetLanguage stringrequired

      Target language

      LLM objectrequired
      LLM parameters for translation.
      Url stringrequired

      LLM interface address to call. Different for different Vendors:

      • When Vendor is DoubaoSeedTranslation: https://ark.cn-beijing.volces.com/api/v3/responses
      • When Vendor is QwenMT: https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions
      ApiKey stringrequired

      The parameter used for authentication by the LLM service provider.

      Model stringrequired

      The LLM model name to call, required. Different vendors support different models.

      Params object

      LLM or application platform request parameters, optional.

      • For OpenAI protocol, you can configure parameters such as max_tokens, temperature, stream, etc.
      • For other application platforms, you can fill in the parameters supported by the platform.
      SubtitleType integer

      Subtitle delivery type, default is 0.
      0: No delivery;
      1: Recognition results only (e.g., when a user says "你好" in Chinese, delivers "你好");
      2: Translation results only (e.g., when a user says "你好" in Chinese, delivers the translated "Hello");
      3: Both recognition and translation results (e.g., when a user says "你好" in Chinese, delivers both "你好" and "Hello").
      If you need to display subtitles on the client, it is recommended to set this to 1, 2, or 3. For details, see Display Subtitles.

      CallbackConfig object
      Callback configuration.
      HostTag string

      You can set up to two callback addresses to monitor events that occur during the task. Use this parameter to specify which callback address the task's events should be routed to. Contact ZEGO technical support to configure the specific tag values in advance. Currently, Custom1 and Custom2 are supported.

    Responses​

    Task started successfully
    Schema
      Code integer

      Return code, 0 means success, other values mean failure. For details, refer to Return Codes.

      Message string

      Request result description

      RequestId string

      Request ID

      Data object
      TaskId string

      Unique identifier of the task.

    Previous

    Return Codes

    Next

    Stop Cloud Real-Time ASR

    On this page

    Back to top