ID_OCR_MAX is an ID Verification solution that uses the Qwen-VL large model to detect and recognize various types of ID certificates from around the world.
In the Indonesia region, this API operation supports only the recognition of Indonesian ID cards (document type code IDN01001).
Endpoint
API operation: DocOcrMax
Description: Uses the Qwen-VL large model to recognize the document type and extract information from an ID certificate image, and returns the relevant fields.
Request method: POST
Transport protocol: HTTPS
QPS limit: This API is subject to an exclusive queries per second (QPS) limit. For more information, see ID Verification server-side API QPS limits.
Endpoint:
NoteAn internal network is a private communication network between Alibaba Cloud products in the same region. If your application server is deployed in an Alibaba Cloud region, you can use the internal endpoint to access the ID Verification service. This provides a more secure and stable network connection.
Singapore
Public endpoint:
cloudauth-intl.ap-southeast-1.aliyuncs.comInternal same-region endpoint:
cloudauth-intl-vpc.ap-southeast-1.aliyuncs.com
Indonesia
Public endpoint:
cloudauth-intl.ap-southeast-5.aliyuncs.comInternal same-region endpoint:
cloudauth-intl-vpc.ap-southeast-5.aliyuncs.com
Online debugging and integration
Before you debug and integrate, read the Use OpenAPI Explorer to debug and integrate server-side API operations document. It explains how to call API operations on the OpenAPI platform and how to obtain the SDK and its sample code.
You can run this API operation in OpenAPI Explorer to debug it. You can also generate SDK sample code for this API operation.
Image format requirements
To ensure stable model performance, the uploaded image must meet all the following requirements:
Image format: JPG, JPEG, or PNG.
Image size: The recommended size is 50 KB to 100 KB. The maximum size is 10 MB.
NoteWhen you pass parameters in base64 format, ensure that the original image size does not exceed 6 MB to stay within the 10 MB data transmission limit.
Image resolution: The image height and width must be between 200 pixels and 8192 pixels. The recommended resolution is 480 × 640 pixels (height × width).
Image quality recommendations:
The four corners of the ID card image must be complete. Avoid obstructing the document when holding it, because this can affect detection.
The content of the ID card image must be clear, unobstructed, and have no overexposed areas. The ID card image must not be tilted or inverted.
The ID card image should occupy more than 60% of the total image area. A smaller card area reduces the recognition accuracy.
We recommend using photos of real documents.
Request parameters
You can provide the image using one of the following two parameters:
IdOcrPictureBase64
IdOcrPictureUrl
Name | Type | Required | Description | Example |
ProductCode | String | Yes | The product solution to use. Set the value to ID_OCR_MAX. | ID_OCR_MAX |
SceneCode | String | No | A custom authentication scenario ID. You can use this ID to query related records in the console. It can be a combination of letters, digits, and underscores, up to 10 characters long. | 1234567890 |
MerchantBizId | String | Yes | A custom unique business ID. Use it to locate and troubleshoot issues later. It can be a combination of letters and digits, up to 32 characters long. Make sure that the ID is unique. | e0c34a77f5ac40a5aa5e6ed20c35**** |
MerchantUserId | String | No | A custom user ID or another identifier for a specific user, such as a mobile number or email address. Desensitize the value of this field in advance. For example, use hashing. | 123456789 |
IdOcrPictureBase64 | String | No | The Base64-encoded image of the ID certificate. If you use IdOcrPictureBase64 to pass the ID certificate image, check the image size. Do not pass an oversized image. | base64 |
IdOcrPictureUrl | String | No | The URL of the ID certificate image. The URL must be an HTTP or HTTPS link accessible over the public network. | https://***.jpg |
OcrModel | String | No | The optical character recognition (OCR) recognition mode: Note Modes 1 and 2 do not support document quality inspection or anti-spoofing detection. For more information, see Differences between recognition modes.
| 0 |
DocType | String | No | Note
| GLB03001 |
DocPage | String | No | The page to recognize.
Important If the page detected by the model does not match the specified page, Subcode 213 is returned. | 01 |
IdSpoof | String | No | Specifies whether to enable the document anti-spoofing feature:
| F |
IdThreshold | String | No | The custom threshold mode for OCR quality inspection:
| 3 |
OcrValueStandard | String | No | Specifies whether to return additional OCR fields in a standardized format:
Note For information about the supported document types and standardized fields, see Field standardization. | 0 |
Authorize | String | No | Specifies whether to enable verification with an authoritative data source to enhance document anti-spoofing capabilities.
Note
| T |
Differences between recognition modes
OCR recognition mode | Input differences | Output differences | Recommended scenarios |
0: Specified document recognition mode |
| This mode outputs detailed fields from the document page. For more information, see Document OCR fields. | Recommended for online admission scenarios for ID Verification. This mode automatically verifies whether the document type is permitted for the business and outputs detailed fields from the document face. |
1: Automatic document classification mode Important This recognition mode is not supported in the Indonesia region. |
| This mode outputs only the detected document type code. For more information, see Global ID certificate codes. | Recommended for offline business analysis scenarios. You can use this mode for clustering and cleansing of historical ID certificate images. |
2: Automatic document classification and general recognition mode Important This recognition mode is not supported in the Indonesia region. |
| This mode outputs the document type and key standardized fields from the document face. For more information, see the table of general recognition fields. Note This mode supports the output of only common fields across different document types. | Recommended for offline scenarios that require assisted manual review. This mode recognizes the document type and key fields to improve review efficiency. |
Field standardization
To handle the diversity of global ID certificates, ID Verification provides a field standardization feature. This feature supports standardized output formats for the following key document fields.
Supported document types
Currently, only the following three certificate types from South Korea are supported. Support for additional certificate types will be added based on demand.
South Korea | KOR01001 | 01 | |
KOR02001 | 01 | ||
KOR03001 | 01 |
Supported standardized fields
Field | Type | Description |
surname_s | String | The surname in Latin letters. |
given_name_s | String | The given name in Latin letters. |
sex_s | String | The sex, F or M. |
date_of_birth_s | String | The date of birth in the yyyy-mm-dd format. |
date_of_expiry_s | String | The expiration date of the document in the yyyy-mm-dd format. |
Response parameters
Name | Type | Description | Example | |
HTTP Status Code | Integer | The HTTP status code. | 200 | |
HTTP Body | RequestId | String | The request ID. | 130A2C10-B9EE-4D84-88E3-5384F******** |
Result.TransactionId | String | The unique ID for the entire authentication process. | hksb7ba1b28130d24e015d694******** | |
Code | String | Success | ||
Message | String | A detailed description of the returned Code. | success | |
Result.Passed | String | Indicates whether the recognition is complete. | Y | |
Result.SubCode | String | 200 | ||
Result.ExtIdInfo | String |
|
| |
ExtIdInfo
Name | Type | Description | Example |
ocrIdInfo | String | The OCR field information of the ID certificate. Note If the ID certificate OCR process fails, this field is empty. | |
spoofInfo | String | The result of anti-counterfeiting detection, including the risk identification result and risk type: Note Card detection is enabled only when you set IdSpoof = Y in the Initialize operation. Otherwise, spoofResult returns N by default, and spoofType is empty.
| |
ocrDocType | String | Note This field is returned only when OcrModel is set to 1 or 2. | CHN01 |
ocrStandardData | String | The standardized OCR field information of the ID certificate. Note This field is returned only if the document type is supported and | |
Return codes
HTTP status code | Code | Message |
200 | Success | Request successful. |
400 | MissingParameter | Parameter cannot be empty. |
InvalidParameter | Invalid parameter. | |
401 | UnqualifiedPhoto | The uploaded image is unreadable or the image resolution does not meet requirements. We recommend changing the image. Ensure the photo is clear, has normal exposure, is complete without occlusion, and has no significant angle deviation. |
DataDuplication | Both Base64-encoded image and URL image address are provided. Choose only one of these parameters. | |
ToolargeImage | The image size is too large. We recommend compressing the image or changing the upload method. | |
DownloadTimeout | URL image download timeout. | |
403 | Forbidden.RAMUserAccessDenied | You need to grant the RAM user the AliyunAntCloudAuthFullAccess permission. For more information, see Authorize RAM users to access the service. |
Forbidden.AccountAccessDenied | Ensure that you have activated ID verification and your account has no overdue payment. | |
Throttling.Api | API request is blocked due to throttling. | |
500 | InternalError | Internal system error. Provide feedback to engineers for troubleshooting. |
ResultObject.SubCode error codes
Subcode is returned based on different product solutions or integration methods. For details, refer to the following table.
Applicable solution | Error code | Is authentication record billed | Description and suggested reasons |
General | 200 | Yes | Authentication passed. |
| 211 | Yes | The quality or resolution of the certificate image does not meet requirements, or the image itself is incomplete. Ensure the photo of the certificate's portrait side is clear, has normal exposure, is complete without occlusion, and has no significant angle deviation. |
| 212 | Yes | Certificate anti-counterfeiting detection indicates risk. There may be high-risk operations such as rephotographing, tampering, or photocopying. |
| 213 | Yes | The specified certificate type was not detected (recognition mode) or the certificate type could not be identified (classification mode). We recommend uploading a clear, complete certificate image with normal angle. |