What is this?

The UMBC AI Gateway provides access to AI models from multiple providers (including Claude, GPT, Gemini, locally-hosted models, and more) through a centralized, university-managed platform. You can interact with these models in two ways:

OpenWebUI — a chat interface in your browser (no setup required)
API key (OpenAI-compatible) — for use in your own scripts, tools, or applications

Your access is managed through an API key tied to a team. Each team has a shared budget that refreshes on a set schedule (e.g., monthly). Once the team budget is reached, requests will be paused until the next refresh. You can learn more about these ideas on the AI Gateway Teams and Budgets page.

To learn more about these and other GenAI tools offered by UMBC, please see the descriptive entires in the UMBC GenAI Software Catalog .

Using OpenWebUI (Chat Interface)

This is the easiest way to get started — no technical setup needed.

Submit an RT Ticket via the “DoIT-AI Tools” Form
Once you are provided access, go to https://openwebui.aws.genai.umbc.edu/
Log in with your normal UMBC credentials via the “Continue with UMBC SSO” button
Select a model from the dropdown at the top-left of the chat
Start typing your prompt

That's it. Your usage is tracked automatically against your team's budget.

For more tips and tricks regarding Open WebUI, see the “Open WebUI Basic Usage” page and it’s children to the left.

Using Your API Key Directly

If you've been given an API key (starts with `sk-`), you can use it with any tool that supports the OpenAI-compatible API format.

Base URL (Gateway)

https://gateway.aws.genai.umbc.edu

Example: curl

curl https://gateway.aws.genai.umbc.edu/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR-KEY-HERE" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_NAME",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Example: Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="sk-YOUR-KEY-HERE",
    base_url="https://gateway.aws.genai.umbc.edu/v1",
)

response = client.chat.completions.create(
    model="MODEL_NAME",
    messages=[{"role": "user", "content": "Hello!"}],
)

print(response.choices[0].message.content)

Replace `MODEL_NAME` with the model you've been authorized to use.

A listing of models available to your key should have been provided alongside your API Key in a Box Note.

Checking Your Usage

curl -s -X GET "https://gateway.aws.genai.umbc.edu/key/info?key=sk-YOUR-KEY-HERE" \                                                                                                       
    -H "Authorization: Bearer sk-YOUR-KEY-HERE"

Replace “sk-YOUR-KEY-HERE” with your API Key.

Look for the `spend` field in the response.

Tips for Managing Usage

Shorter prompts cost less. Be concise when possible
Choose the right model. Smaller/faster models are cheaper for simple tasks
Avoid unnecessary retries. Each request counts toward your budget, even if you don't use the response

Getting Help

Submit a descriptive RT Ticket via the “DoIT-AI Tools” Form if you need …

A new or replacement API key
A budget increase
Access to additional models
Help with integration into your workflow

Notice an issue? Please submit a descriptive RT Ticket via the “DoIT-AI Tools” Form. We’re dedicated to making these pages clear, concise, and useful!

AI Gateway & Open WebUI