Skip to content

Getting Started: Configuration & Authentication

This guide helps you set up your client environment and authenticate with the PowerMind ComputeAI platform.

1. Prerequisites

Before you start, make sure you have:

  1. Network Access: The network IP or domain name of your PowerMind ComputeAI Controller API.
  2. CLI Executable: The pm-compute command-line utility available on your machine.
  3. Active Email Address: If you need to register a new account.

2. CLI Setup & API Endpoint Configuration

The pm-compute CLI needs to know where the central Controller API is running.

2.1 Set the API URL

To configure the base API endpoint, use the config set command:

pm-compute config set --api-url https://<computeai-controller-host>

Tip

If you are working in a laboratory or local development environment that uses self-signed SSL/TLS certificates, add the --insecure flag to disable certificate verification:

pm-compute config set --api-url https://<computeai-controller-host> --insecure

2.2 Verify Configuration

To check your active configuration:

pm-compute config view

To test if the Controller API is reachable:

pm-compute health
If successful, you will see a message: API is healthy along with the version number.

3. Account Provisioning (Signing Up)

If your administrator has enabled public registration, you can create a new account directly from the CLI:

pm-compute signup -u <username> -p '<password>' -e <email>

Password Requirements:

  • At least 12 characters.
  • Must contain at least one uppercase letter.
  • Must contain at least one lowercase letter.
  • Must contain at least one number.

4. User Authentication (Logging In)

To sign in and create an active session on your terminal:

pm-compute login -u <username> -p '<password>'

If you do not specify the username or password options, the CLI will prompt you securely:

pm-compute login
# Username: alice
# Password: *************

Upon successful login, a JSON token is saved locally to authorize subsequent commands.

5. Session Management

ComputeAI supports multi-user session management on a single machine. Sessions are scoped to your local terminal.

5.1 Check Current User

To see which account is currently active:

pm-compute whoami

5.2 List Active Sessions

To see all users who have active session files saved on this machine:

pm-compute sessions
This lists usernames, roles, and the last login timestamps.

5.3 Switch Sessions

If you have logged in with multiple accounts, you can switch the active session:

pm-compute switch <USERNAME>
You will be prompted to enter that user's password to re-authenticate.

5.4 Execute a Single Command as Another User

You can temporarily bypass the active terminal session to run a single command as another logged-in user:

pm-compute --user <USERNAME> jobs

6. Logging Out

To terminate your session and clear local authorization tokens for your active terminal:

pm-compute logout
This deletes the terminal-specific session file, preventing unauthorized access from that console.

Next Steps

Once you have configured the CLI and logged in, you can start preparing and executing ML training jobs. For a complete step-by-step tutorial, see the End-to-End MLOps Workflow. For detailed syntax of all CLI commands, see the CLI Command Reference Manual.