Platform For AI:Create and manage workstations

Create a workstation

A workstation is an LLM intelligent routing service bound to a group that shares its name. All services in the workstation belong to that group.

1. Log in to the PAI console. In the left navigation pane, choose Model Deployment > InferX.

2. If this is your first time, click Create Workstation on the welcome page.

   If a workstation already exists, click the current workstation name in the upper-right corner and create a new one.

3. Configure the following parameters:

   Parameter	Description
   Workstation Name	Custom name. Use lowercase letters, digits, and underscores. The name must start with a lowercase letter and cannot be changed after creation.
   Resource Type	Select Public Resource Group (pay-as-you-go, suitable for quick testing) or Resource Quota (pre-purchased resources, suitable for production).
   Deployment Resources	Resources allocated to the intelligent routing component that receives and forwards requests. Default: 2 replicas, 4 CPU cores, and 8 GB of memory. For production workloads, increase the replica count to improve routing stability.
   Scheduler Resource Configuration	Resources allocated to the scheduling management component. Default: 2 CPU cores and 4 GB of memory.
   Gateway Selection	Select Shared Gateway or Dedicated Gateway: Shared Gateway: Free, with no SLA guarantee. Suitable for development and testing. Dedicated Gateway: Requires a paid subscription. Supports public network allowlists, VPC allowlists, and cross-region deployment. Recommended for production. For details, see Use a dedicated gateway.
   VPC Configuration	Optional. Configure the VPC, vSwitch, and security group for the workstation. Note A VPC is required when using Redis.
   Redis Configuration	Optional. When enabled, user data and usage statistics are persisted in Alibaba Cloud Tair (Redis-compatible). Without Redis, data is stored in local memory only and is lost when the service stops. We recommend that you configure Redis for production environments. Choose a configuration method based on whether you already have a Redis instance: Create a new Instance: Recommended when you need to create a new instance. The system automatically creates a Tair instance and configures the allowlist. After you click Create a Tair instance, the Tair instance creation progress appears. Do not refresh the page during this process. If you refresh the page, go to the Tair console to complete the remaining steps. After creation, the Redis address, Redis username (default: default), and Redis password are displayed. Store this information securely. Select Existing Instance: If you have an existing Tair or Redis instance in your account, select it to avoid creating a duplicate instance and incurring additional costs. Manual Input: The Redis instance and the workstation must be in the same VPC; otherwise, the connection will fail.

4. Review the total cost at the bottom, then click Create Workstation.

View calling information

After the workstation is created, click Call Information in the upper-right corner to view the Base URL and API key for calling inference services.

• Base URL: Provides both public and VPC access addresses. The VPC address offers lower latency and incurs no public traffic costs. Use the VPC address when the caller and the workstation are in the same VPC.

• API Key: Used for request authentication. Include the API key in HTTP request headers as Authorization: Bearer <API_KEY>. Click the copy icon to get the full key.

Delete a workstation

1. Click Details in the upper-right corner to open the workstation details panel.

2. At the bottom of the details panel, click Delete Workstation, then click OK.

Upgrade the engine version

When a new version of the LLM intelligent routing engine is released, you can upgrade in one of the following ways without recreating the workstation.

LLM intelligent routing engine release notes

The following table describes the changes in each version of the LLM intelligent routing engine.