Call ChatWithDesensitize API to Enable Privacy-Safe AI Chat - Data Management

Chat API: Data Masking with Data Management Service (DMS)

Operation description

A Chat API with data masking capabilities powered by Data Management (DMS).

Try it now

Try this API in OpenAPI Explorer, no manual signing needed. Successful calls auto-generate SDK code matching your parameters. Download it with built-in credential security for local usage.

Test

RAM authorization

The table below describes the authorization required to call this API. You can define it in a Resource Access Management (RAM) policy. The table's columns are detailed below:

Action: The actions can be used in the Action element of RAM permission policy statements to grant permissions to perform the operation.
API: The API that you can call to perform the action.
Access level: The predefined level of access granted for each API. Valid values: create, list, get, update, and delete.
Resource type: The type of the resource that supports authorization to perform the action. It indicates if the action supports resource-level permission. The specified resource must be compatible with the action. Otherwise, the policy will be ineffective.
- For APIs with resource-level permissions, required resource types are marked with an asterisk (*). Specify the corresponding Alibaba Cloud Resource Name (ARN) in the Resource element of the policy.
- For APIs without resource-level permissions, it is shown as All Resources. Use an asterisk (*) in the Resource element of the policy.
Condition key: The condition keys defined by the service. The key allows for granular control, applying to either actions alone or actions associated with specific resources. In addition to service-specific condition keys, Alibaba Cloud provides a set of common condition keys applicable across all RAM-supported services.
Dependent action: The dependent actions required to run the action. To complete the action, the RAM user or the RAM role must have the permissions to perform all dependent actions.

Action

Access level

Resource type

Condition key

Dependent action

dms:ChatWithDesensitize

none

*All Resource

*

None

Request syntax

POST / HTTP/1.1

Request parameters

Parameter	Type	Required	Description	Example
InstanceId	integer	Yes	The ID of the instance containing the desensitization rule to be applied. Obtain this ID by calling the ListInstances or GetInstance API.	123***
Messages	array	No	A chronological list of message objects that provides context for the Large Language Model (LLM).	[ { "content": "你好", "role": "user" } ]
	any	No	A message object in the conversation. The standard format is as follows: { "content": "$MessageContent", "role": "$MessageRole: system, user, or assistant" }	{ "content": "你好", "role": "user" }
Model	string	No	The ID of the model to use. Supported models include text-only Large Language Models (LLMs) from the Qwen series, such as commercial and open-source versions. Multimodal models are not supported.	qwen-plus
Stop	array	No	A list of stop words.	["\n"]
	string	No	A stop word. Generation stops immediately when the model outputs a string listed in the `Stop` array.	\n
NeedDesensitization	boolean	No	Specifies whether to enable data desensitization. The default value is `false`.	false
DesensitizationRule	string	No	The category of the desensitization rule. This parameter is required when `NeedDesensitization` is set to `true`.	UserInfo
MaxTokens	integer	No	The maximum number of tokens to generate. Generation stops once the output reaches this limit.	256
PresencePenalty	string	No	Controls token repetition in the generated text. Valid range: `[-2.0, 2.0]`. Positive values reduce repetition, and negative values increase it.	0.0
ResponseFormat	string	No	The format of the response content. Valid values are `text` for a plain text response and `json_object` for a standard JSON string.	text
Seed	integer	No	A random number seed. Use this parameter to ensure reproducible results for the same input and parameters. Valid range: `[0, 2³¹−1]`.	1
EnableThinking	boolean	No	Specifies whether to enable the thinking mode when you use a hybrid thinking model.	true
ThinkingBudget	integer	No	The maximum number of tokens for the thinking process.	256
Temperature	string	No	The sampling temperature, which controls the diversity of the generated text. A higher temperature results in more diverse text, while a lower temperature produces more deterministic text. Valid range: `[0, 2.0)`.	1
TopLogprobs	integer	No	Specifies the number of most likely candidate tokens for which to return logprobs at each generation step. Valid range: `[0, 5]`.	1
TopK	integer	No	Specifies the number of candidate tokens to sample from during generation. A higher value results in more random output, while a lower value produces more deterministic output. This feature is disabled if the value is set to `null` or greater than `100`.	10
TopP	string	No	The probability threshold for nucleus sampling, which controls text diversity. Higher values result in more diverse text. Valid range: `(0, 1.0]`.	0.5
XDashScopeDataInspection	string	No	Specifies whether to perform additional violation detection on input and output content in addition to the built-in content safety features of the Qwen API.	{}
SearchOptions	object	No	The strategy for web search.	{}
	string	No	Key-value pairs for the web search strategy.	{}
ModalitiesList	array	No	Specifies the desired output modalities. This parameter applies only to the Qwen-Omni model.	["text","audio"]
	string	No	Specifies a single output modality.	text
AudioJson	string	No	The timbre and format settings for the audio output. This parameter applies only to the Qwen-Omni model and when the `ModalitiesList` parameter is set to `["text","audio"]`.	{}
EnableCodeInterpreter	boolean	No	Specifies whether to enable the Code Interpreter feature. This parameter takes effect only when the `model` is `qwen3-max-preview` and `EnableThinking` is `true`.	false
Logprobs	boolean	No	Specifies whether to return the log probabilities (logprobs) for each output token.	false
VlHighResolutionImages	boolean	No	Specifies whether to increase the input image pixel limit to the equivalent of 16,384 tokens.	false
EnableSearch	boolean	No	Specifies whether to enable web search.	false

Response elements

Element	Type	Description	Example
	object	The response body.
RequestId	string	The ID of the request.	0C1CB646-1DE4-4AD0-B4A4-7D47DD52E931
ErrorCode	string	The error code.	UnknownError
ErrorMessage	string	The error message.	UnknownError
Success	boolean	Indicates whether the request was successful. Valid values: true: The request was successful. false: The request failed.	true
Data	object	The returned data.
Created	string	The Unix timestamp (in seconds) when the request was created.	1763710100
Model	string	The model used for this request.	qwen-plus
Choices	array<object>	An array of candidate responses generated by the model.
	array<object>	A candidate response generated by the model.
FinishReason	string	The reason why the model stopped generating the response. Valid values:	stop
Message	object	The message generated by the model.
Content	string	The content of the model's response.	你好呀！
ReasoningContent	string	The in-depth reasoning content from the model.	嗯，用户发了个“你好”，看起来是想打招呼...
Role	string	The role of the message.	system
Logprobs	object	Log probability information for the model's output tokens.	{}
	any	Detailed token information.	{}
Usage	object	Token usage information for this request.
CompletionTokens	string	The number of tokens in the model's output.	10
PromptTokens	string	The number of tokens in the input.	9
TotalTokens	string	The total number of tokens consumed.	19
PromptTokensDetails	object	A fine-grained breakdown of input tokens.	{}
	string	A fine-grained token category.	{}
CompletionTokensDetails	object	A fine-grained breakdown of output tokens when you use the Qwen-VL model.	{}
	string	A fine-grained token category.	{}
Message	string	The error message. This parameter is returned only if the `StatusCode` is not `200`.	InvalidParameter
StatusCode	string	The status code. A value of `200` indicates a successful call. Other values indicate an error.	200
Type	string	The error type.	invalid_request_error

Examples

Success response

JSON format

{
  "RequestId": "0C1CB646-1DE4-4AD0-B4A4-7D47DD52E931",
  "ErrorCode": "UnknownError",
  "ErrorMessage": "UnknownError",
  "Success": true,
  "Data": {
    "Created": "1763710100",
    "Model": "qwen-plus",
    "Choices": [
      {
        "FinishReason": "stop",
        "Message": {
          "Content": "你好呀！",
          "ReasoningContent": "嗯，用户发了个“你好”，看起来是想打招呼...",
          "Role": "system"
        },
        "Logprobs": {
          "key": "{}"
        }
      }
    ],
    "Usage": {
      "CompletionTokens": "10",
      "PromptTokens": "9",
      "TotalTokens": "19",
      "PromptTokensDetails": {
        "key": "{}"
      },
      "CompletionTokensDetails": {
        "key": "{}"
      }
    },
    "Message": "InvalidParameter",
    "StatusCode": "200",
    "Type": "invalid_request_error"
  }
}

Error codes

See Error Codes for a complete list.

Release notes

See Release Notes for a complete list.