All Products
Search
Document Center

Platform For AI:User guide for new users

Last Updated:Dec 16, 2025

This topic shows you how to get started with Platform for AI (PAI) and introduces typical AI development workflows on the platform.

1. Activate the product

Log on to the PAI console, select a region in the upper-left corner, and then click Activate Now. After you activate the service, a default workspace is automatically created.

To prevent activation failures due to insufficient AliyunPAIFullAccess permissions, use your Alibaba Cloud account to activate the service.

2. Quick Start

  1. Choose a functional module based on its description.

  2. Follow its quick start guide to learn the basics through a simple example.

  3. Read the module's user guide to explore more features and best practices.

    Click the link in the Functional module column to go to the user guide for that module.

Functional module

Description

Quick Start

Data Science Workshop (DSW)

Provides a cloud-based IDE for AI development. Developers familiar with Notebooks or VSCode can quickly start model development.

DSW Quick Start

Machine Learning Designer

Offers over 140 built-in algorithm components for low-code, visual modeling using a drag-and-drop interface.

Designer Quick Start

Deep Learning Containers (DLC)

Allows you to quickly create distributed or standalone training tasks without manually provisioning machines or configuring environments, mirroring the local experience of running training scripts.

DLC Quick Start

Elastic Algorithm Service (EAS)

Deploys trained models as an online inference service with simple configuration.

EAS Quick Start

Model Gallery

Integrates DLC and EAS to train and deploy open-source large language models efficiently with zero code.

Model Gallery Quick Start

3. Get help

The FAQ documents contain common questions and solutions from developers using PAI. If you encounter issues with modules like DSW or EAS, refer to the corresponding FAQ document:

DSW FAQ

EAS FAQ

DLC FAQ

Model Gallery FAQ

iTAG FAQ

Designer FAQ

Billing FAQ

PAI ArtLab FAQ

4. Typical AI development process

PAI covers the entire AI development process, from data preparation and model training to model deployment. The following sections describe two typical AI development processes.

Cloud-native AI development

image

Step

Description

References

Dataset management in PAI centrally manages local, cloud, and public datasets as data sources for model training.

Dataset management

DSW provides a cloud-based IDE for AI development. Developers familiar with Notebook or VSCode can quickly start model development.

Create a DSW instance

Image management in PAI centrally manages official public images and your custom images. An image provides the runtime environment for your code.

Image management

After developing and testing your model code in DSW, use DLC to run training tasks for higher efficiency and cost savings.

Create a DLC training task

PAI supports mounting file systems (such as NAS and OSS) and Git repositories, which simplifies specifying data and code when you submit a task.

Code management

Model management centrally manages trained models, which can be directly deployed by using EAS.

Model management

After a model is trained, use EAS to quickly deploy the model as an online inference service.

Deploy an EAS model service

AI and big data development

image

Step

Description

References

If your data is stored in MaxCompute, first use DataWorks for data pre-processing. You can then reference the MaxCompute table as the training data source in PAI.

General data development

Designer offers over 140 built-in algorithm components for low-code, visual modeling using a drag-and-drop interface.

Visual Modeling Designer

Use DataWorks to configure and run periodically scheduled tasks.

Node scheduling configuration

Task management in Designer records execution details for experiments and custom tasks, which simplifies task comparison and analysis.

Task management

Model management centrally manages trained models, which can be directly deployed by using EAS.

Register and manage models

After a model is trained, use EAS to quickly deploy the model as an online inference service.

Deploy an EAS model service

FAQ

Q: The Activate Now button is grayed out and a message appears: "The current account does not have permission to activate PAI. Contact the owner of your Alibaba Cloud account." What should I do?

Solutions:

  • Use your Alibaba Cloud account to complete the activation.

  • If you are using a Resource Access Management (RAM) user, you must add the AliyunPAIFullAccess system policy to the RAM user. Note: This policy grants extensive permissions. The Alibaba Cloud account administrator should assess the security risks before granting it.

Q: What should I do if the error message  Create order error: message is Your account balance is less than 0. Please top up your account and try to purchase again. productRequestId is *** is displayed when I try to activate PAI??

Go to the User Center page to view your bills and pay any outstanding fees. Then, try activating PAI again.