Integration overview

Versions

Alibaba Cloud OpenAPI uses version numbers to manage the API lists for major versions of cloud products. For example, Vision Intelligence Platform currently supports the API list for version 2024-01-11. The string 2024-01-11 is an API version number, not a date. The list shows the latest public API data, and the version number does not mean that the data has remained unchanged since that date.

Online debugging

The OpenAPI portal provides features such as API debugging for Vision AI Platform. Before making a call, familiarize yourself with the versions, endpoints, and integration methods provided by Vision AI Platform.

On the API debugging page in the OpenAPI Explorer, the left pane lists the available APIs. In the central configuration area, you can select a service region, such as China (Shanghai), and fill in required parameters like ImageURL by uploading a file or entering a URL. Click Initiate Call to start online debugging. The right-side panel provides reference information, including Documentation, Call Result, SDK Example, and CLI Example.

Debugging entry point

An example entry point for API debugging is: https://next.api.aliyun.com/api/facebody/2019-12-30/AddFaceEntity

Endpoints

Select an endpoint based on the region where your resources are located for the lowest latency. For example, for Optical Character Recognition (OCR) in the China (Shanghai) region, the public endpoint is ocr.cn-shanghai.aliyuncs.com and the VPC endpoint is ocr-vpc.cn-shanghai.aliyuncs.com.

• Public endpoints are globally accessible.

• VPC endpoints are accessible only from within the corresponding Alibaba Cloud region and VPC network. VPC endpoints offer the following advantages:

  • High security: Access is restricted to your VPC, providing enhanced security and privacy.

  • Faster response: Because they operate on the internal VPC network, VPC endpoints offer faster response times than public endpoints and avoid issues like public network latency.

  • Lower cost: VPC endpoints use internal network communication.

For more information, see Endpoints.

Supported identities

Logging in to the OpenAPI Explorer with your Alibaba Cloud account grants full API access, posing a high security risk. For better security, we strongly recommend using a RAM user with the minimum necessary permissions for API calls and daily O&M. For more information, see Create a RAM user to manage Vision AI Platform.

More information

• Identities, credentials, and authorization

• Throttling and quota management

Integration methods

• Alibaba Cloud SDK

  • Alibaba Cloud provides SDKs for multiple programming languages, including Java, C#, Go, Python, Node.js/TypeScript, PHP, and C++. You can integrate an SDK and call APIs directly by calling its methods. The SDKs handle request signatures, timeouts, and retries, and they return structured response objects to simplify development. For more information about Alibaba Cloud SDKs, see Alibaba Cloud SDK.

  • You can use Alibaba Cloud SDKs to call Vision AI Platform APIs. For information about supported languages and how to install dependencies, see the following topics:

    • Vision AI Platform - Image Generation

    • Vision AI Platform - Image Recognition

    • Vision AI Platform - Object Detection

    • Vision AI Platform - Product Understanding

    • Vision AI Platform - Optical Character Recognition (OCR)

    • Vision AI Platform - Content Moderation

    • Vision AI Platform - Image Matting and Segmentation

    • Vision AI Platform - Face and Body

    • Vision AI Platform - Image Analysis and Processing

    • Vision AI Platform - Visual Search

    • Vision AI Platform - Video Understanding

    • Vision AI Platform - Video Segmentation

    • Vision AI Platform - Video Generation

• Manual API integration

  If you make API calls by using native HTTP, you must implement the signature algorithm and build the requests yourself. For more information about the signature mechanism, see API overview and Request signature.

Usage notes

If a call to the Vision AI Platform API returns an error, you need to check whether the input request parameters and their values are correct based on the returned error code. For more information, see the common error codes for each feature in the API Reference. For example, see Common error codes for object detection for Object detection.

You can also record the RequestID from the response or the error message from the SDK and use the Alibaba Cloud OpenAPI Diagnostic Platform to diagnose the issue.

FAQ

• To learn more about Vision AI Platform, see What is Vision AI Platform?.

• Vision AI Platform provides cost-effective, accessible, and easy-to-use AI capabilities. The platform delivers a wide range of visual AI capabilities across multiple domains, including general vision, images, videos, and object recognition. These capabilities are organized into 13 categories with hundreds of AI features, including Face and Body, Optical Character Recognition (OCR), Product Understanding, Content Moderation, Image Recognition, Image Generation, Image Matting and Segmentation, Visual Search, Image Analysis and Processing, Object Detection, Video Understanding, Video Generation, and Video Segmentation. The platform is continuously updated with more visual AI capabilities. All capabilities are available through APIs, providing a standard way to integrate and use them. For the latest information, see the Vision AI Platform product page.

• In addition to the Java SDK, languages such as C#, Go, Python, Node.js/TypeScript, PHP, and C++ are also supported. For more information, see SDK Reference.