Skip to content

Latest commit

 

History

History
130 lines (79 loc) · 5.14 KB

File metadata and controls

130 lines (79 loc) · 5.14 KB

Getting Started

Your first steps into the CUNY CAIL Sandbox — from logging in to having your first conversation with an AI model.


What You'll Find Here

The Sandbox gives you a shared AI platform where you can create custom agents, ground them in your course materials, and share configurations with students or colleagues. It runs on CUNY infrastructure with zero-retention agreements: prompts and responses are not stored or used for model training. Chat history persists in Open WebUI and is accessible only to you and system administrators.


First Login

What You Need

  • CUNY credentials (the same login you use for other CUNY services)
  • A web browser (Chrome, Firefox, Safari, or Edge)
  • About 10 minutes to explore

Step by Step

  1. Navigate to chat.ailab.gc.cuny.edu

    • Bookmark this page. You'll be back often.
  2. Click "Sign In" at the top right

    • The Sandbox uses CUNY's authentication system, so you'll be redirected to the standard CUNY login page
  3. Enter your CUNY credentials and complete any two-factor authentication if prompted

    • Same process as accessing CUNY email or Blackboard
  4. You'll land on the main chat interface

    • Clean layout. Chat input at the bottom. Model selector at the top. Sidebar on the left with workspace features.

First login walkthrough

Voila! You're in. Now let's have a conversation.


Your First Conversation

Try This

In the chat input at the bottom of the screen, type something simple:

"Explain the concept of scaffolding in education, suitable for graduate students in a teaching methods course."

Press Enter (or click the send button). The AI model responds.

The model name at the top of the screen shows which one you're talking to. Click it to switch. Different models have different strengths: writing, reasoning, image understanding, speed. You can switch mid-conversation.


Exploring the Interface

For terminology and interface layout, see Sandbox Basics.


Customizing Your Experience

Settings You Might Want to Adjust

  1. Click your profile avatar (top right)
  2. Select Settings
  3. You'll see several tabs. Here's what matters for getting started:

General Tab:

  • Default Model — Choose which AI model loads when you start a new chat
    • Recommendation: Start with a general-purpose model like DeepSeek V3.2 or Kimi K2.5
  • Interface Theme — Light, dark, or auto (matches your system settings)
    • Dark mode reduces eye strain during long writing sessions
  • Chat Input Mode — Desktop defaults to Enter to send, Shift+Enter for new line
    • If you prefer multiline input, flip this setting

Account Tab:

  • Profile Picture — Optional, but helps when collaborating with colleagues
  • Display Name — How your name appears to others when you share models or knowledge bases

Settings panel with key options highlighted

Save and Return to Chat

Once you've adjusted your preferences:

  1. Scroll to the bottom of the Settings panel
  2. Click Save
    • Your settings persist across sessions
  3. Click the X or press Escape to close Settings
  4. You're back to the chat interface, ready to work

What's Next?

To create custom AI models, knowledge bases, and tool configurations, request Workspace access by emailing ailab@cuny.edu.

Once you have access, head to the Design & Test section to:

  • Create custom models with specific behaviors
  • Build knowledge bases from your course materials
  • Bind tools that extend what the AI can do

In the meantime, explore: try different models (click the model name at the top), upload a document and ask questions about it, or experiment with different prompts.


Tips for Getting Started

Start small. You don't need to build complex configurations on day one. Chat with the default models. See what works. See what doesn't. Let that guide what you build later.

Save interesting conversations. Click the chat title in the sidebar to rename it. This makes it easier to find later when you want to revisit a useful interaction.

Don't worry about breaking things. The Sandbox is a shared tool, but your personal workspace is yours. Create models. Delete them. Experiment. You can't break the instance for others.

Ask for help. If you get stuck, reach out to the AI Lab team or check the community channels. CUNY faculty and researchers are building with this platform every day, and the community is generous with support.


Additional Resources


Continue to Sandbox Basics →