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

Text Moderation 2.0 upgrades multi-language model capabilities with automatic language detection and expanded language support. It provides moderation policies and a labeling system designed for international business. This topic describes the multi-language service for Text Moderation 2.0 and how to use it.

Features

Compared to the multi-language moderation service of Text Moderation 1.0, Text Moderation 2.0 uses a separate policy and labeling system designed for international business. It also provides additional features to simplify integration and assist manual review.

Feature	Text Moderation 2.0	Text Moderation 1.0
Language support	Supports 38 languages.	Supports 18 languages.
Moderation capabilities	Runs multiple models in parallel, combining language and regional characteristics for more precise policies.	Uses a single model incorporating language characteristics. Its policy balances precision and recall.
Labeling system	Uses an internationalized labeling system with additional labels for profanity and regional content, and supports multiple risk labels and subcategory labels.	Uses the labeling system designed for Chinese-language contexts and supports only a single risk label.
API behavior	Does not require you to specify the input language—it automatically detects the language. After moderation, the service returns the detected language type and an English translation of the content to assist manual review.	Requires you to specify the input language. No translation is returned.

Supported languages

Text Moderation Enhanced Edition supports 38 languages.

Language	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
Uyghur	Uyghur	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
Greek	Greek	el
Romanian	Romanian	ro
Irish	Irish	ga

Internationalization labels

The multi-language service of Text Moderation Enhanced Edition uses an internationalized labeling system. If content contains multiple risk categories, the service returns multiple labels simultaneously. The label categories include but are not limited to the following:

Type

Category

Level-1 labels (labels)

violence: violence
contraband: contraband
sexuality: pornography
profanity: profanity and abuse
pullinTraffic: advertising content
regional: regional hostility
C_customized: custom library hit

Subcategory labels (riskTips)

Subcategory labels are returned in the xxx_yyy format. For example: contraband_Drugs.

Subcategory labels help identify specific violation reasons and allow you to handle content based on regional compliance requirements. For a complete list of subcategory label types, contact our support team.

Billing methods

Text Moderation Enhanced Edition supports pay-as-you-go and resource package deduction billing.

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 actual usage. No fees are charged if you do not call the service.

Type	Service	Unit price
Advanced text moderation (text_advanced)	Multi-language detection for international business: comment_multilingual_pro	CNY 15 per 10,000 calls

Resource package deduction

If you have a high moderation volume or relatively consistent moderation needs, we recommend purchasing a resource package. Larger packages offer greater discounts. Multiple packages can be purchased and stacked. For more information, see Purchase a resource package for Content Moderation Enhanced Edition.

This resource package covers Content Moderation Enhanced Edition usage and cannot be shared with Content Security 1.0 traffic packages. The deduction factors are as follows:

Type	Service	Deduction factor
Advanced text moderation (text_advanced)	Multi-language detection for international business: comment_multilingual_pro	The deduction factor is 2, meaning each successful API call deducts 2 units from the resource package capacity. Note For example, if your resource package has a capacity of 10 calls, one successful API call deducts 2 calls, leaving a remaining balance of 8 calls.

Integration

Step 1: Activate the service

Go to service activation to activate the Text Moderation 2.0 service.

After you activate the Text Moderation 2.0 service, the default billing method is pay-as-you-go. You are billed daily based on actual usage. No fees are charged if you do not call the service. After you integrate with the API, the system automatically generates bills based on your usage. For more information, see Billing.You can also purchase a resource package. Resource packages offer tiered discounts compared to pay-as-you-go pricing and are suitable for users with predictable or high usage volumes.

Step 2: Grant permissions to a RAM user

Before you integrate the SDK or call the API, you must grant the required permissions to a RAM user. You can create an AccessKey pair for your Alibaba Cloud account or for a RAM user. All API calls require an AccessKey for authentication. For instructions, 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: Integrate the SDK

The following table lists the supported regions. For the Text Moderation Enhanced Edition SDK, see Integration guide.

Region	Public endpoint	VPC endpoint
China (Shanghai)	green-cip.cn-shanghai.aliyuncs.com	green-cip-vpc.cn-shanghai.aliyuncs.com
China (Beijing)	green-cip.cn-beijing.aliyuncs.com	green-cip-vpc.cn-beijing.aliyuncs.com
China (Hangzhou)	green-cip.cn-hangzhou.aliyuncs.com	green-cip-vpc.cn-hangzhou.aliyuncs.com
China (Shenzhen)	green-cip.cn-shenzhen.aliyuncs.com	green-cip-vpc.cn-shenzhen.aliyuncs.com
China (Chengdu)	green-cip.cn-chengdu.aliyuncs.com	N/A
Singapore	green-cip.ap-southeast-1.aliyuncs.com	green-cip-vpc.ap-southeast-1.aliyuncs.com
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	N/A
UK (London)	green-cip.eu-west-1.aliyuncs.com	N/A

Note

If you need SDK sample code in other programming languages, use OpenAPI Explorer to debug API operations online. The tool automatically generates SDK sample code for the corresponding API.

API

Usage

Endpoint: https://green-cip.{region}.aliyuncs.com.

Call this operation to create a text content moderation task. For information about how to construct an HTTP request, see Request structure. You can also use a pre-constructed HTTP request. For more information, see Access guide.

Billing:
This is a paid operation. Only requests that return an HTTP status code of 200 are metered and billed. Requests that return other error codes are not charged. For more information about billing, see Billing methods.

QPS limit

The per-user QPS limit for this operation is 100 calls per second. Exceeding this limit triggers throttling, which may affect your business. Plan your API calls accordingly.

Request parameters

Parameter	Type	Required	Example	Description
Service	String	Yes	comment_multilingual_pro	The moderation service type. Set the value to comment_multilingual_pro for multi-language detection for international business.
ServiceParameters	JSONString	Yes		The parameter set required by the moderation service, in JSON string format. For descriptions of each field, see ServiceParameters.

Table 1. ServiceParameters

Parameter	Type	Required	Example	Description
content	String	Yes	testing content	The text content to moderate. Maximum length: 600 characters.
dataId	String	No	text0424****	The data ID corresponding to the object to moderate. Contains uppercase and lowercase letters, digits, underscores (_), hyphens (-), and periods (.). Maximum length: 64 characters. Use this to uniquely identify your business data.
accountId	String	No	10123****	The account ID that uniquely identifies an account.
deviceId	String	No	20240307****	The device ID that uniquely identifies a device.
deviceToken	String	No	MzQvo1d7scyZ3tl_RcJ******	The device token obtained through the risk control SDK.

Response parameters

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

Table 2.Data

Parameter	Type	Example	Description
Labels	String	sexual_content	The risk labels. Multiple labels are separated by commas (,). Valid values: violence: violence contraband: contraband sexuality: sexual content profanity: profanity pullinTraffic: promotional content regional: regional hostility C_customized: custom library hit
Reason	String	{"customizedWords": "HXXXXX","detectedLanguage": "en","riskTips": "sexuality_Suggestive","riskWords": "pxxxxy","translatedContent": "pxxxxy sxxxx, HXXXXX","customizedLibs": "ZXXXXX"}	A JSON string for locating the reason, which includes the following fields: riskLevel: The risk level, returned based on the system-recommended risk level. Valid values: high: high risk medium: medium risk low: low risk Note High-risk content: take action immediately. Medium-risk content: send for manual review. Low-risk content: handle only when high recall is needed. Otherwise, treat it the same as content with no risk detected. riskTips: Subcategory labels. riskWords: The risk content that was matched. customizedWords: The custom keywords that were matched. customizedLibs: The name of the custom library that was matched. translatedContent: The translated text content. Returned only after the text translation feature is enabled. Note The text translation feature is currently supported only in the Singapore region. Configure it by managing detection rules in the console. Enabling this feature incurs additional costs. For billing details, see Billing methods.
AccountId	String	10123****	The account ID. Note This field is returned only if you pass `accountId` in the request parameters.
DeviceId	String	20240307****	The device ID. Note This field is returned only if you pass `deviceId` in the request parameters.

Examples

Request example

{
    "Service": "comment_multilingual_pro",
    "ServiceParameters": {
        "content": "testing content",
        "accountId": "10123****",
        "dataId": "text0424****"
    }
}

Response examples:

If both custom keywords and risk content are detected:

{
    "Code": 200,
    "Data": {
        "DataId": "text0424****",
        "Labels": "C_customized,profanity",
        "Reason": "{\"riskLevel\":\"high\",\"customizedWords\":\"HXXXXX\",\"detectedLanguage\":\"en\",\"riskTips\":\"sexuality_Suggestive\",\"riskWords\":\"pxxxxy\",\"translatedContent\":\"pxxxxy sxxxx，HXXXXX\",\"customizedLibs\":\"ZXXXXX\"}"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

If only custom keywords are detected:

{
    "Code": 200,
    "Data": {
        "DataId": "text0424****",
        "Labels": "C_customized",
        "Reason": "{\"riskLevel\":\"high\",\"customizedWords\":\"HXXXXX\",\"detectedLanguage\":\"en\",\"translatedContent\":\"HXXXXX\",\"customizedLibs\":\"ZXXXXX\"}"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

If only risk content is detected:

{
    "Code": 200,
    "Data": {
        "DataId": "text0424****",
        "Labels": "sexuality",
        "Reason": "{\"riskLevel\":\"high\",\"detectedLanguage\":\"ar\",\"riskTips\":\"sexuality_Suggestive\",\"riskWords\":\"pxxxxy\",\"translatedContent\":\"pxxxxy sxxxx\"}"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

If neither custom keywords nor risk content are detected:

{
    "Code": 200,
    "Data": {
        "DataId": "text0424****",
        "Labels": "",
        "Reason": "{\"detectedLanguage\":\"en\",\"translatedContent\":\"AXXXXXX\"}"
    },
    "Message": "OK",
    "RequestId": "AAAAAA-BBBB-CCCCC-DDDD-EEEEEEEE****"
}

Error codes

Code	Status message	Description
200	OK	The request is successful.
400	BAD_REQUEST	Invalid request. Check your request parameters.
407	NOT_SUPPORT	The specified language is not recognized or is unsupported.
408	PERMISSION_DENY	Possible causes: the account is not authorized, has overdue payments, the service is not activated, or the account is disabled.
500	GENERAL_ERROR	An internal server error occurred. Retry the request. If the error persists, contact Support and Services.
581	TIMEOUT	The request timed out. Retry the request. If the error persists, contact Support and Services.
588	EXCEED_QUOTA	The request frequency exceeds the QPS limit.