Text moderation enhanced edition 2.0 multilingual plus-AI Guardrails(AI Guardrails)-阿里云帮助中心

Text Moderation Enhanced Edition features an upgraded multilingual model that can automatically detect languages and supports more language types. Tailored for international business, the service also provides moderation policies and a labeling system. This topic describes the features and usage of the multilingual service for Text Moderation Enhanced Edition.

Features

Compared to the multilingual service in Text Moderation 1.0, Text Moderation 2.0 uses a separate policy and labeling system for global businesses. It also offers more features to simplify usage and aid manual review.

Item	Text Moderation 2.0	Text Moderation 1.0
Multilingual support	Supports 38 languages.	Supports 18 languages.
Moderation capability	Uses multiple models in parallel. Policies are more precise, tailored to specific languages and regions.	Uses a single model. Policies balance precision and recall based on language features.
Labeling system	Uses an internationalized labeling system that includes labels for profanity and region-specific content. Supports multiple risk labels and fine-grained labels.	Uses a labeling system designed for Chinese content and supports only a single risk label.
Detection scope	You can configure all detection scopes in the console, enabling or disabling them as needed. Each scope maps one-to-one with a detection result.	Supports general detection scopes, which do not map one-to-one with detection results.
API features	Provides automatic detection without requiring you to specify the input language. After moderation, the API returns the language type and a translated English version of the text to aid manual review.	Requires you to specify the input language and does not return translated content.

Supported languages

Text moderation Enhanced Edition supports 38 languages.

Language	English name	Language code
English	English	en
Simplified Chinese	Simplified Chinese	zh
Traditional Chinese	Traditional Chinese	zh-tw
Indonesian	Indonesian	id
Malay	Malay	ms
Thai	Thai	th
Vietnamese	Vietnamese	vi
Tagalog	Tagalog	tl
Hindi	Hindi	hi
Arabic	Arabic	ar
Turkish	Turkish	tr
French	French	fr
German	German	de
Russian	Russian	ru
Portuguese	Portuguese	pt
Spanish	Spanish	es
Italian	Italian	it
Dutch	Dutch	nl
Polish	Polish	pl
Japanese	Japanese	ja
Korean	Korean	ko
Urdu	Urdu	ur
Uighur	Uighur	ug
Bengali	Bengali	bn
Persian	Persian	fa
Swedish	Swedish	sv
Danish	Danish	da
Norwegian	Norwegian	no
Icelandic	Icelandic	is
Finnish	Finnish	fi
Belarusian	Belarusian	be
Lithuanian	Lithuanian	lt
Czech	Czech	cs
Slovak	Slovak	sk
Hungarian	Hungarian	hu
Modern Greek	Modern Greek	el
Romanian	Romanian	ro
Irish	Irish	ga

Internationalized labels

The Content ModerationEnhanced Edition Multilingual PLUS service uses an internationalized label system. If content poses multiple risks, the service returns multiple labels. These labels include:

Label	Confidence score	Description
pornographic_adult	0–100. A higher score indicates greater confidence.	Suspected pornographic content
sexual_terms	0–100. A higher score indicates greater confidence.	Suspected sexual health content
sexual_suggestive	0–100. A higher score indicates greater confidence.	Suspected vulgar content
sexual_orientation	0–100. A higher score indicates greater confidence.	Suspected sexual orientation content
regional_cn	0–100. A higher score indicates greater confidence.	Suspected domestic political content
regional_illegal	0–100. A higher score indicates greater confidence.	Suspected illegal political content
regional_controversial	0–100. A higher score indicates greater confidence.	Suspected political controversy
regional_racism	0–100. A higher score indicates greater confidence.	Suspected racism
violent_extremist	0–100. A higher score indicates greater confidence.	Suspected extremist organizations
violent_incidents	0–100. A higher score indicates greater confidence.	Suspected extremist content
violent_weapons	0–100. A higher score indicates greater confidence.	Suspected weapons and ammunition
violence_unscList	0–100. A higher score indicates greater confidence.	United Nations Security Council Consolidated List
contraband_drug	0–100. A higher score indicates greater confidence.	Suspected drug-related content
contraband_gambling	0–100. A higher score indicates greater confidence.	Suspected gambling-related content
inappropriate_ethics	0–100. A higher score indicates greater confidence.	Suspected unethical content
inappropriate_profanity	0–100. A higher score indicates greater confidence.	Suspected abusive or insulting content
inappropriate_oral	0–100. A higher score indicates greater confidence.	Suspected vulgar language
inappropriate_religion	0–100. A higher score indicates greater confidence.	Suspected religious profanity
pt_to_contact	0–100. A higher score indicates greater confidence.	Suspected traffic diversion using contact information
pt_to_sites	0–100. A higher score indicates greater confidence.	Suspected off-site traffic diversion
customized	0–100. A higher score indicates greater confidence.	Matched a keyword in a custom keyword library

Billing

The text moderation Enhanced Edition service supports pay-as-you-go and resource plan offset.

Pay-as-you-go

After you activate the text moderation Enhanced Edition service, the default billing method is pay-as-you-go. You are billed daily based on your actual usage. If you do not use the service, you will not be charged.

Moderation type	Service	Unit price
Advanced text moderation (text_advanced)	Multilingual detection for international business: comment_multilingual_pro_cb	CNY 15 per 10,000 API calls

Resource plan offset

For large or consistent moderation needs, we recommend purchasing a resource plan. Larger plans offer greater discounts. You can purchase and use multiple resource plans. For more information, see Purchase a resource plan for Content Moderation Enhanced Edition.

This resource plan offsets usage for Content Moderation Enhanced Edition and is not compatible with Content Moderation 1.0 usage plans. The following table describes the offset factors.

Moderation type	Service	Offset factor
Advanced text moderation (text_advanced)	Multilingual detection for international business: comment_multilingual_pro_cb	The offset factor is 2. Each successful API call deducts 2 API calls from your resource plan's usage quota. Note For example, if your resource plan has a usage quota of 10 API calls, one successful API call deducts 2 API calls from the plan, leaving a balance of 8 API calls.

Integration

Step 1: Activate the service

To activate the Text Moderation Enhanced Edition service, visit Activate Service.

After you activate the Text Moderation Enhanced Edition service, the default billing method is pay-as-you-go. You are billed daily based on your actual usage. You are not charged if you do not call the service. After you integrate the API, the system automatically bills you based on your usage. For more information, see Billing. You can also purchase a pay-as-you-go resource plan, which offers tiered discounts and is suitable for users with predictable or high usage volumes.

Step 2: Grant permissions to a RAM User

Before integrating the SDK or API, grant permissions to a RAM User. You can create an AccessKey for your Alibaba Cloud account or a RAM User. An AccessKey authenticates your identity when you call Alibaba Cloud APIs. To learn how to get an AccessKey, see Obtain an AccessKey.

Grant permissions to a RAM user

Log on to the RAM console using your Alibaba Cloud account.
Create a RAM user. For details, see Create a RAM user.
Grant the AliyunYundunGreenWebFullAccess system policy to the RAM user. This policy grants full access to Content Moderation. For details, see Manage RAM user permissions.
The RAM user can now call the Content Moderation API.

Step 3: Install and integrate the SDK

The service is available in the following regions. To integrate the SDK for Text Moderation Enhanced Edition, see the Integration Guide.

Region	Public endpoint	Internal endpoint
Singapore	green-cip.ap-southeast-1.aliyuncs.com	green-cip-vpc.ap-southeast-1.aliyuncs.com
UK (London)	green-cip.eu-west-1.aliyuncs.com	Not available
US (Virginia)	green-cip.us-east-1.aliyuncs.com	green-cip-vpc.us-east-1.aliyuncs.com
US (Silicon Valley)	green-cip.us-west-1.aliyuncs.com	Not available
Germany (Frankfurt)	green-cip.eu-central-1.aliyuncs.com	green-cip-vpc.eu-central-1.aliyuncs.com

Note

Manage configurations for the UK (London) region in the Singapore region console, and for the US (Silicon Valley) region in the US (Virginia) region console.

API

Usage

API operation: TextModerationPlus

Use this operation to create a text content moderation task. To learn how to construct an HTTP request, see Request structure. You can also use a prebuilt HTTP request. For more information, see Integration guide.

Billing:
This is a paid operation. You are charged only for successful requests that return a 200 status code. Requests that return other status codes are not billed. For more information about billing, see Billing.

QPS limit

The QPS limit for this operation is 100 requests per second per user. API calls that exceed this limit are throttled, which may affect your business. Plan your calls accordingly.

Request parameters

Parameter	Type	Required	Example	Description
Service	String	Yes	comment_multilingual_pro_cb	The type of the moderation service. Set the value to comment_multilingual_pro_cb for multilingual moderation for international services.
ServiceParameters	JSONString	Yes		The set of parameters for the moderation service, provided as a JSON string. For details about the parameters, see ServiceParameters.

Table 1. ServiceParameters

Parameter

Type

Required

Example

Description

content

String

Yes

testing content

The text content to moderate. The content can be up to 600 characters long.

dataId

String

text0424****

The data ID of the object to moderate.

The ID can contain uppercase and lowercase letters, digits, underscores (_), hyphens (-), and periods (.). The ID can be up to 64 characters long and can be used to uniquely identify your business data.

Response parameters

Parameter	Type	Example	Description
Code	Integer	200	The status code. For more information, see Code details.
Data	JSONObject		The moderation result data. For more information, see Data.
Message	String	OK	The response message.
RequestId	String	AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****	The request ID.

Table 2. Data

Parameter	Type	Example	Description
Result	JSONArray		The moderation results, including risk labels, confidence scores, and other details. For more information, see Result.
DataId	String	text0424****	The data ID of the moderated object. Note This parameter is returned only if `dataId` is specified in the request.
RiskLevel	String	high	The risk level, determined by the configured risk score thresholds. Valid values: `high`: High risk. If the content matches a keyword in a custom keyword library, the risk level defaults to high. `medium`: Medium risk. `low`: Low risk. `none`: No risk detected. Note We recommend handling high-risk content directly and manually reviewing medium-risk content. Process low-risk content only if a high recall rate is required; otherwise, it can be treated as risk-free. You can configure risk score thresholds in the Content Moderation console.
TranslatedContent	String	Translated text	The translated text content. Returned only after enabling the text translation feature. Note The text translation feature is currently supported only in the Singapore region. Configure it under detection rule management in the console. Enabling this feature incurs additional fees. For billing details, see Billing.
DetectedLanguage	String	en	The detected language.

Table 3. Result

Parameter	Type	Example	Description
Label	String	political_xxx	The moderation label. A single piece of content may receive multiple labels and confidence scores. For a list of supported labels, see Internationalized labels.
Confidence	Float	81.22	The confidence score, ranging from 0 to 100. The value is precise to two decimal places. Some labels do not have a confidence score.
Riskwords	String	AA,BB,CC	The detected sensitive words. Multiple sensitive words are separated by commas. This parameter is not returned for some labels.
CustomizedHit	JSONArray	[{"LibName":"...","Keywords":"..."}]	Returned if the content matches a keyword in a custom keyword library (when `Label` is `customized`). This parameter provides the name of the custom library and the matched custom keywords. For more information, see CustomizedHit.
Description	String	Suspected pornographic content	A description of the `Label` field. Important This field provides a human-readable explanation of the `Label` and is subject to change. For automated processing, base your logic on the `Label` field, not this `Description` field.

Table 4. CustomizedHit

Parameter	Type	Example	Description
LibName	String	Custom library 1	The name of the custom keyword library.
Keywords	String	Custom keyword 1,Custom keyword 2	The matched custom keywords. Multiple keywords are separated by commas.

Examples

Request example

{
    "Service": "comment_detection_pro_cb",
    "ServiceParameters": {
        "content": "testing content",
        "dataId": "text0424****"
    }
}

Response examples

System policy match:

{
    "Code": 200,
    "Data": {
        "Result": [
            {
                "Label": "political_entity",
                "Description": "Suspected political entity",
                "Confidence": 100.0,
                "RiskWords": "Word A,Word B,Word C"
            },
            {
                "Label": "political_figure",
                "Description": "Suspected political figure",
                "Confidence": 100.0,
                "RiskWords": "Word A,Word B,Word C"
            }
        ],
        "RiskLevel": "high",
        "DetectedLanguage": "en",
        "TranslatedContent": "Translated content",
        "DataId": "text0424****"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

Custom keyword library match:

{
    "Code": 200,
    "Data": {
        "Result": [
            {
                "Description": "Matched a custom keyword library",
                "CustomizedHit": [
                    {
                        "LibName": "Custom keyword library name 1",
                        "KeyWords": "Custom keyword"
                    }
                ],
                "Confidence": 100,
                "Label": "customized"
            }
        ],
        "RiskLevel": "high",
        "DataId": "text0424****"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

Code

Code	Status code	Description
200	OK	The request was successful.
400	BAD_REQUEST	Invalid request. Check your request parameters.
407	NOT_SUPPORT	The language is not recognized or supported.
408	PERMISSION_DENY	Permission denied. This can occur if your account is not authorized, has an overdue payment, or if the service is inactive or blocked.
500	GENERAL_ERROR	A server-side error occurred. Retry the request. If the error persists, contact us through online support.
581	TIMEOUT	The request timed out. Retry the request. If the error persists, contact us through online support.
588	EXCEED_QUOTA	The request rate exceeds the QPS limit.