Call /green/text/scan to moderate text - Content Moderation

Usage notes

Operation: /green/text/scan

Submit text moderation tasks through this endpoint. For details on constructing HTTP requests, see Request syntax. To use an SDK instead, see SDK overview.

By default, this operation moderates Chinese text. To moderate text in other languages, contact your account manager. Supported languages: English, French, German, Indonesian, Malaysian, Portuguese, Spanish, Thai, Vietnamese, Japanese, Arabic, Filipino, Hindi, Turkish, Russian, Italian, and Dutch.

Billing

Calling this operation incurs charges. For details, see AI Guardrails Pricing.

QPS limit

This operation supports up to 100 calls per second per account. Exceeding this limit triggers throttling, which may affect your business. Factor this limit into your implementation.

Request parameters

Parameter	Type	Required	Example	Description
bizType	String	No	default	The business scenario. Create business scenarios in the AI Guardrails console. For details, see Customize policies for machine-assisted moderation.
scenes	StringArray	Yes	["antispam"]	The moderation scenario. Set to `antispam` for text moderation.
tasks	JSONArray	Yes		The list of text entries to moderate. Contains up to 100 elements, each structured as a task object (see the following table). To submit more than 100 entries per request, contact your account manager to increase the concurrency limit.

task

Parameter	Type	Required	Example	Description
clientInfo	JSONObject	No		Client information. For details, see the "Common request parameters" section of Common parameters. The task-level `clientInfo` takes priority over the global `clientInfo` parameter.
dataId	String	No	cfd33235-71a4-468b-8137-a5ffe323\\\\	A unique identifier for the moderated object. Accepts letters, digits, underscores (\_), hyphens (-), and periods (.). Maximum length: 128 characters.
content	String	Yes	We are a micro-lending company that provides an unsecured loan on the same day with reliable, fast, convenient, and door-to-door services when you apply for a loan on your mobile phone.	The text to moderate. Maximum length: 10,000 characters, including punctuation.

Response parameters

Parameter	Type	Example	Description
code	Integer	200	HTTP status code. For details, see Common error codes.
msg	String	OK	The response message.
dataId	String	cfd33235-71a4-468b-8137-a5ffe323\\\\	The ID of the moderated object. Returned only if `dataId` was set in the request.
taskId	String	txt6HB8NQoEbU@5fosnj2xVEM-1t\\\\	The moderation task ID.
content	String	We are a micro-lending company that provides an unsecured loan on the same day with reliable, fast, convenient, and door-to-door services when you apply for a loan on your mobile phone.	The original text submitted in the request.
filteredContent	String	We are a \\\\ company that provides an \\\* loan on the same day with reliable, fast, convenient, and \\\\ services when you apply for a loan on your mobile phone.	The text with matched custom terms replaced by asterisks (\*). Customize terms through the CreateKeywordLib API or the AI Guardrails console. For details, see Manage custom text libraries.
results	JSONArray		The moderation results. On success (HTTP 200), returns an array of result objects (see the following table).

result

Parameter	Type	Example	Description
scene	String	antispam	The moderation scenario specified in the request.
suggestion	String	block	The recommended action. Valid values: `pass` (text is normal), `review` (text requires human review), `block` (text contains violations and can be deleted or blocked).
label	String	porn	The content category. Valid values: `normal` (normal content), `spam` (junk content), `ad` (advertising content), `politics` (political content), `terrorism` (terrorist content), `abuse` (abusive content), `porn` (pornographic content), `flood` (excessive junk content), `contraband` (prohibited content), `meaningless` (meaningless content), `harmful` (harmful content, including content related to mammonism, wealth flaunting, blind worshiping of idols, disruptive emotion, and negative enticing), `customized` (custom content, such as a custom term).
rate	Float	99.91	The confidence score. Valid values: 0 to 100. Higher values indicate greater confidence. When `suggestion` is `pass`, a higher score means the content is more likely normal. When `suggestion` is `review` or `block`, a higher score means the content is more likely to contain violations. Use the `suggestion`, `label`, and `sublabel` parameters to determine whether content contains violations. The `sublabel` parameter is returned by specific operations.
extras	JSONObject	{"userId":"xxx"}	Additional information (extended parameter).
details	JSONArray		Details of the risky content detected. A text entry can match multiple risk categories. Each element is a detail object (see the following table).

detail

Parameter	Type	Example	Description
label	String	porn	The risk category of matched content. Valid values: `spam` (junk content), `ad` (advertising content), `politics` (political content), `terrorism` (terrorist content), `abuse` (abusive content), `porn` (pornographic content), `flood` (excessive junk content), `contraband` (prohibited content), `meaningless` (meaningless content), `harmful` (harmful content, including content related to mammonism, wealth flaunting, blind worshiping of idols, disruptive emotion, and negative enticing), `customized` (custom content, such as a custom term).
contexts	JSONArray		Context information for the matched risky content. Each element is a context object (see the following table).

context

Parameter	Type	Example	Description
context	String	door-to-door	The term matched in the text. Returned only for term-based matches, not for algorithmic model matches.
positions	JSONArray	[{"startPos":1, "endPos":10}]	The position of the matched term in the original text.
libName	String	Name of your custom text library	The name of the custom text library. Returned only when the text matches a term in a custom text library.
libCode	String	123456	The code of the custom text library. Returned only when the text matches a term in a custom text library.
ruleType	String	ip	The behavior rule type. Returned only when the text matches a behavior rule. Valid values: `user_id`, `ip`, `umid`, `content`, `similar_content`, `imei`, `imsi`.

Examples

Sample request

http(s)://[Endpoint]/green/text/scan
&<Common request parameters>

{
  "scenes": [
    "antispam"
  ],
  "tasks": [
    {
      "dataId": "cfd33235-71a4-468b-8137-a5ffe323****",
      "content": "We are a micro-lending company that provides an unsecured loan on the same day with reliable, fast, convenient, and door-to-door services when you apply for a loan on your mobile phone."
    }
  ]
}

Sample response

{
  "code": 200,
  "data": [
    {
      "code": 200,
      "content": "We are a micro-lending company that provides an unsecured loan on the same day with reliable, fast, convenient, and door-to-door services when you apply for a loan on your mobile phone.",
      "dataId": "cfd33235-71a4-468b-8137-a5ffe323****",
      "filteredContent": "We are a **** company that provides an *** loan on the same day with reliable, fast, convenient, and **** services when you apply for a loan on your mobile phone.",
      "msg": "OK",
      "results": [
        {
          "details": [
            {
              "contexts": [
                {
                  "context": "unsecured",
                  "positions": [
                    {
                      "endPos": 19,
                      "startPos": 16
                    }
                  ]
                },
                {
                  "context": "micro-lending",
                  "positions": [
                    {
                      "endPos": 6,
                      "startPos": 2
                    }
                  ]
                }
              ],
              "label": "spam"
            },
            {
              "contexts": [
                {
                  "context": "door-to-door",
                  "libCode": "123456",
                  "libName": "Name of your custom text library",
                  "positions": [
                    {
                      "endPos": 34,
                      "startPos": 30
                    }
                  ]
                }
              ],
              "label": "porn"
            }
          ],
          "label": "porn",
          "rate": 99.91,
          "scene": "antispam",
          "suggestion": "block"
        }
      ],
      "taskId": "txt6HB8NQoEbU@5fosnj2xVEM-1t****"
    }
  ],
  "msg": "OK",
  "requestId": "25711794-BF6D-4F32-A735-09CA21197D32"
}