Get started with Anyscale

Check your docs version

This version of the Anyscale docs is deprecated. Go to the latest version for up to date information.

1. Register an organization

• Join Anyscale: Anyscale is invitation-only. To join, request an invite here.
• Receive Your Invite Code: After approval, look for an invite code.
• Register Your Organization: Use the invite code to register at https://console.anyscale.com.

Understanding organizations in Anyscale

• What is an Organization?: It's a group of users, typically from the same company. All billing is managed at the organization level.
• Creating an Organization: This happens when a new user, not part of an existing organization, registers on Anyscale. This user becomes the organization's owner.
• Adding Team Members: As an owner, inviting others via email is possible. Remember, an email address can only link to one organization.

2. Set up your local environment

AWS

Step 1: Install the Anyscale CLI and Python client package

pip install -U anyscale

Step 2: Authenticate the CLI

To access Anyscale's services, you need to obtain a token that verifies your identity. Running the following command fetches and updates the token in the local credential file ~/.anyscale/credentials.json:

anyscale login

If necessary, log in to the Anyscale console to complete authentication.

GCP

Step 1: Install the Anyscale CLI and Python client package

pip install -U "anyscale[gcp]"

Step 2: Authenticate the CLI

To access Anyscale's services, you need to obtain a token that verifies your identity. Running the following command fetches and updates the token in the local credential file ~/.anyscale/credentials.json:

anyscale login

If necessary, log in to the Anyscale console to complete authentication.

3. Deploy an Anyscale Cloud

AWS

Step 1: Configure AWS credentials

Ensure you have set up AWS credentials on the local machine. Refer to the AWS configuration guide for initial setup instructions. Run this command to quickly set and view the credentials:

aws configure

🔍Verify AWS Credentials Permissions

• Confirm the credentials can launch EC2 instances in the desired AWS region.
• Check permissions for managing VPC, subnets, Security Group, IAM, S3, and EFS. Find detailed requirements in the Anyscale AWS Management Guide.

Step 2: Create your Anyscale Cloud

To set up your Anyscale Cloud with default settings, run the following command:

anyscale cloud setup

For troubleshooting and advanced customization, follow the complete Anyscale Cloud on AWS guide.

Step 3: Verify the Anyscale Cloud

To verify that the Anyscale Cloud satisfies the minimum resource requirements, run the following command to launch a test Workspace and Service:

anyscale cloud verify --name NAME_OF_CLOUD --functional-verify workspace,service

GCP

Step 1: Configure your GCP Account

• Follow the Google Cloud instructions for installing the gcloud CLI.
• Create a GCP project for Anyscale to launch clusters in.

🔍Verify GCP Credentials Permissions

• Confirm you are the owner of the GCP project in which Anyscale operates, by opening IAM & Admin in the Google Cloud Platform Console.

Step 2: Create the Anyscale Cloud

To set up the Anyscale Cloud with default settings, run the following command:

anyscale cloud setup

For troubleshooting and advanced customization, follow the complete Anyscale Clouds on GCP guide.

Step 3: Verify the Anyscale Cloud

To verify that the Anyscale Cloud satisfies the minimum resource requirements, run the following command to launch a test Workspace and Service:

anyscale cloud verify --name NAME_OF_CLOUD --functional-verify workspace,service

4. Add team members

• View Organization Members: To see your team, click the user icon and select My organization.
• Invite Teammates: To add more people, click "Invite teammates" and enter their email addresses, separated by commas. This sends out invitation links to join.
• Manage Roles: Assigning roles to team members is possible. Choose between 'owner' for administrator rights and 'collaborator' for standard access.

What's next?

➡️ Anyscale Private Endpoints - Quickly deploy and fine-tune large language models through API endpoints.

➡️ Anyscale AI Applications - Build sophisticated, production-grade AI applications.