All Products
Search
Document Center

Dataphin:Smart applications

Last Updated:Nov 18, 2025

You can connect large AI models, such as Alibaba Cloud Model Studio and DeepSeek, to smart applications. These models help you discover, process, develop, and manage data to reduce manual costs. This topic describes how to configure smart applications.

Prerequisites

You must activate at least one intelligent feature to access the smart application management page.

Permission description

Super administrators, system administrators, operations administrators, and custom global roles with Smart Application Management - Management permissions can configure model invocation and enable or disable applications.

Model invocation configuration

Before enabling smart applications, you need to configure model invocation. This process involves selecting a model service provider, configuring the API key required to invoke the model, and enabling the models you want to use. All models that can be invoked by smart applications are sourced from this configuration.

  1. In the top menu bar of the Dataphin home page, choose Super X > Smart Application Management.

  2. In the navigation pane on the left, choose Smart Application Management > Smart Applications. On the Smart Applications page, click the Model Invocation Configuration button in the upper-right corner. In the Model Invocation Configuration dialog box, select a model service provider. You can select Alibaba Cloud (Model Studio), DeepSeek, or Alibaba Cloud (AI Stack).

  3. Click the Edit icon next to the model service provider. In the dialog box, configure the parameters.

    Parameter

    Description

    API Key (required)

    Enter the API key for invoking the model provided by the specified model service provider. The path to obtain the API key varies for each model provider. For more information, see the operation instructions for each model platform.

    For example, for the Alibaba Cloud Model Studio platform, see Obtain an API key for detailed steps.

    Base URL

    Enter the base URL to use. If you leave this blank, the system uses the default base URL from the official model service provider. The default base URL for Alibaba Cloud (Model Studio) is https://dashscope.aliyuncs.com/compatible-mode/v1. The default for DeepSeek is https://api.deepseek.com/v1.

    After configuration, click Connection Test to test the connectivity of the API key and base URL you entered. If the test fails, you can view the failure reason.

    Model List

    Displays the available models under the specified model service provider. You can choose to enable or disable a model.

    Note
    • Disabling a model also disables the smart applications that reference this model. For smart applications that support multiple models, if all added models are disabled, the system automatically disables the smart application.

    • The system determines whether to support thinking mode based on model properties. If thinking mode is supported, you can enable or disable it during conversations. For more information about the supported built-in models, see Built-in models supported by smart applications.

  4. Click OK to complete the model service provider configuration.

Smart application management

On the Smart Applications page, activated smart applications are displayed as cards. The information on each card includes the smart application name, type, model, and a description. For the data analytics assistant, the intelligent analysis collection is also shown. You can also edit, enable, disable, and use the smart applications.

  • Edit: Click More in the upper-right corner of the card and select Edit. In the dialog box, configure the parameters and click OK to save the changes.

    • Model: You can select any model that has been enabled for a configured model service provider. For assistants that support multiple models, you can switch between them during conversations as needed.

    • Limited Result Set: You can configure this setting if data sampling is enabled in Administration > Metadata > Sampling Configuration. You can pass sample data as asset information to the large language model (LLM). This helps the LLM better understand the data content and improves the accuracy of its thinking. You can select Use or Do Not Use. If you select Use, you must first carefully read the Sample Data Transmission Statement. Then, select the I Agree To Transmit Relevant Data Samples To The Large Model Provider Per The Sample Data Transmission Statement checkbox.

  • Enable/Disable Smart Application: This action controls the availability of the smart application. Click More in the upper-right corner of the card and select Enable/Disable Smart Application to enable or disable the application.

  • Use: Before you can use a smart application, you must configure a model for it and ensure the application is enabled. You also need the required permissions to use the application. For more information, see the permission details for each smart application. To use an application, hover your mouse over its card and click Use to navigate to the application's page. Note that the O&M Assistant must be used on a mobile device.

References