tencent cloud


Setting Moderation Policy

Last updated: 2024-01-31 16:44:15


    When you use the content moderation service, you can specify a moderation category through a moderation policy to moderate content in custom scenes. For different file types, CI provides different scenes for your choice, making it easier for you to customize moderation policies that suit your business.
    Currently, the file types and corresponding moderation scenes supported by CI are as follows:
    Supported File Types
    Moderation Scenes
    Specific Moderation Categories
    Image moderation policy, video moderation policy, audio moderation policy, text moderation policy, file moderation policy
    Pornographic content
    Sexually suggestive/vulgar behaviors
    Genital nudity/Sexual behaviors
    Sex toys
    Sexy content
    Pornographic text moderation
    Advertising content
    QR code and barcode recognition
    Logo detection and recognition
    Advertising text moderation
    Other non-compliant content
    Fire, explosion, bloody scene, etc.
    WEBP images involving advertising QR codes cannot be moderated currently.


    Default policy

    Each moderation type has a default moderation policy, which is configured by Tencent Cloud based on your historical moderation conditions. If you have never used the moderation service, the policy developed by algorithm experts based on models for multiple industries will be used by default, which will be suitable for most content security requirements.
    The default policy can be viewed and edited but not deleted.

    Custom policy

    If the default policy cannot meet your business needs, or you have multiple scenes that require different moderation policies, you can create custom policies as needed.
    1. Log in to the CI console.
    2. Click Bucket Management on the left sidebar.
    3. Click the name of the target bucket to enter the configuration page.
    4. Select Content Moderation > Moderation Policy on the left sidebar.
    5. Create a moderation policy based on your business needs. Currently, you can create policies for image moderation, video moderation, audio moderation, text moderation, and file moderation. Take the image moderation policy as an example: i. Click Create Image Moderation Policy and enter the policy name. ii. Select the category you need to moderate under the moderation type, where OCR pornographic text moderation indicates to moderate image content based on OCR.
    iii. Click Save to create the policy.
    6. After the moderation policy is created, the backend will automatically generate a unique Biztype value.
    7. You can view or edit the created moderation policy; however, you cannot change its name and Biztype value.

    Using moderation policy

    After the moderation policy is created, during the configuration of automatic auditing, creation of historical data audit tasks, and invocation of the content audit API interface, you are required to select the corresponding strategy. This allows for auditing to be conducted in accordance with your custom classifications.

    Automatic moderation

    Create an automatic moderation configuration in the console, and you can manually select a moderation policy.

    Historical data moderation

    Create a historical data moderation job in the console, and you can manually select a moderation policy.

    Moderation APIs

    When calling an API for content moderation, you can manually pass in the Biztype value to the API, which you can view in the list of moderation policies in the console. If you don't pass it in, the default policy will be used.
    Contact Us

    Contact our sales team or business advisors to help your business.

    Technical Support

    Open a ticket if you're looking for further assistance. Our Ticket is 7x24 avaliable.

    7x24 Phone Support