To start using UT Spark go to https://spark.utexas.edu.

Logging In

Once you arrive at the login page, click the Sign in With Microsoft button and you will be asked to provide your UT credentials.

UT Spark - Login Page

Provide your Institutional Identifier (IID) (format: <eid>@eid.utexas.edu) such as: your_eid@eid.utexas.edu

Microsoft IID Prompt

The Microsoft Sign in page will redirect you to the UT Authentication page, your IID will be prefilled in the username field. Enter your UT EID password to sign in.

UT Austin Authentication Page

If you have not completed multi-factor authentication with Duo recently, you will be taken to the Duo prompt.

Example Duo Prompt

Using the Chat

After successfully logging in, you will arrive at a similar screen as shown below. You can begin chatting with the default agent after reading and accepting the guidelines. Click either of the “Guidelines” buttons to continue.

UT Spark - Landing after login

Agent Guidelines

Before being able to interact with any UT Spark agent, you will need to read and agree to the UT Spark Guidelines.

You will need to re-accept the guidelines on each new session of UT Spark.

After clicking the “I Agree” button, you will then be able to use the message bar.

UT Spark - Accepting guidelines before using chat feature

Your first prompt

With the message bar enabled, simply type your prompt into the text box. Press Enter or click the paper airplane icon to start the conversation.

UT Spark’s response will appear directly below your message in the conversation window.

You can then follow up, refine your request, or ask new questions.

Here are some example prompts you can try:

“Summarize this PDF in plain language for a non-technical audience.”
“Draft a professional email to my professor requesting an extension on an assignment.”
“Explain the main themes of The Great Gatsby in 5 bullet points for a study guide.”
“Generate three research questions for a paper on climate change and urban policy.”
“Create a step-by-step outline for a 45-minute lesson plan on machine learning basics.”

See our Prompting 101 article for more information on how to optimize your prompts for best results.

UT Spark - Example chat interaction

User Interface Elements

Message Bar

Below is an image of the message bar with each feature highlighted. We’ll go over what each of these does.

UT Spark - Menu Bar

Text Input Area: This is where you will type a message to the agent or chat.

Send Message: When you are ready to send the message you can hit the [Enter] key or click on Send Message.

File Attachment: You may upload files and images depending on agent settings, this option will displayed here to the far left of the message bar.

For currently accepted file types and limits see: https://cloud.wikis.utexas.edu/wiki/spaces/UTAICommons/pages/449446143/UT+Spark+FAQs#Q%3A-What-file-types-can-I-upload%3F

Speak to Text Feature: You may use your microphone to speak a message and it will be displayed as text in the message bar, hit [Enter] or Send Message to send the text.

Note: There is also a Text To Speech option which will be shown in Response Options.

User Prompt Library: Frequently used prompts can be saved here and used across chats and agents, providing quick access to common queries and commands.

Clicking on the User Prompt Library will result in a menu flyout on the right side of the interface, click on + Add Prompt to input your prompt and give it a title as shown below.

User Prompt Library

Response Options

UT Spark - Response Options

Copy to Clipboard: Use Copy to Clipboard to easily copy the response and paste it somewhere else.

Text to Speech: This will read the answer provided, please keep in mind that the voice will be the OS system default for whichever device is being used. This may or may not be able to be changed depending on your operating system.

Regenerate Response: Depending on the Agent being used the option to regenerate the response can be presented, you can do this if the response isn’t to your liking or if there was a timeout or incomplete response.

Regenerate Response

Interaction Feedback

Clicking on one of the feedback options will result in a prompt to provide feedback about the answer.

Provide Positive Feedback

Provide Negative Feedback

Example Feedback Form

UT Spark Agents

What Is an Agent?

An agent is like a smart helper built on top of the AI model. Instead of just answering your questions, an agent can take actions to complete tasks for you.

Think of it this way:

The model = answers your question.
The agent = figures out the steps needed, calls the right tools, and delivers the result.

Example:

Model: “What’s the weather in Austin today?” → gives you the answer if it already knows.
Agent: “What’s the weather in Austin today?” → goes out, checks a live weather API, and then tells you the answer.

So, when you see the word agent related to AI, just remember:
👉 It’s the AI acting like a helper that can plan, look things up, and get work done — not just chat.

Official Agents

Official Agents in UT Spark are created and managed by the UT Spark admin team. Soon, teams and departments will also have access to the Official Agent Toolkit, allowing them to create their own official agents. These agents can be shared broadly with all UT Spark users or restricted to specific teams, departments, or individuals as needed.

Unlike personal agents, official agents are built with a more advanced set of features and tools, enabling richer capabilities and broader integrations.

More details, including how to apply for access to the Official Agent Toolkit, will be shared soon.

Personal Agents

Personal Agents can be created by clicking on + Create a Personal Agent. You can read more about creating personal agents in UT Spark here.

Selecting an Agent

All available agents can be found by clicking the “Agents” tab and selecting “View all Agents” from within the panel. The agent dashboard allows you to choose which agent you would like to start a chat with.

If you wish to search for an agent, you can use the search bar at the top of the dashboard. However, please note that if you click on New Chat (upper right corner) without selecting an agent, you will be sent to the default agent (or whichever agent you most recently used).

UT Spark - Agent selection dashboard

Creating a Personal Agent

If you want to create your own personal agent, click any of the highlighted buttons in the image below.

To chat with any of the available agents, simply click Chat . The Agent interface will then launch.

Pinning Agents

You may pin an agent to refer to it later without having to search for it. Hover over the agent in the “View All Agents” panel or in the RECENT tab if you’ve interacted with it before. You will see an ellipsis with the option to Pin.

Pinning Agent Action

You may find your pinned agents in the left sidebar of the agent menu.

Pinned Agents Tab

Chat history

You may manage your chat history with UT Spark, in the Chats menu on the left side of the interface you will see a list of your chat history. Clicking on the ellipsis will result in a context menu with a few choices.