Authentication

API requests use up to two keys, both passed in the JSON request body (not headers).

Project API key (api_key)

Identifies and authorizes access to your project.

• Format: uc_ followed by 32 hex characters.
• Where to get it: open https://chat.illinois.edu/<project-name>/api (also reachable from the Materials page) and click Generate API Key. Only project admins and owners can access this page.
• One key at a time: each project has at most one active key. Rotate it with the Rotate button or delete it if needed. Rotating invalidates the previous key immediately.
• The UI shows pre-filled curl and language-specific snippets with your key already inserted.

Treat API keys as secrets

Anyone with your key can chat against your project (and spend your provider credits if you pass a provider key alongside it). Store keys in environment variables or a secrets manager, never in client-side code or version control.

LLM provider key (openai_key)

When using a commercial model, supply your own provider key per request:

Your provider key is never stored

Provider keys are passed through per-request for security and simplicity. That way you control costs and your key is never retained server-side.

The openai_key parameter is optional when using free NCSA-hosted models (e.g. llama3.1:70b) or retrieval_only requests.

Errors

Status	Meaning
401 Unauthorized	Invalid or missing api_key.
403 Forbidden	The key is valid but lacks permission for the requested project.

Example

import os
import requests

response = requests.post(
    "https://chat.illinois.edu/api/chat-api/chat",
    json={
        "model": "gpt-4o-mini",
        "messages": [{"role": "user", "content": "Hello!"}],
        "course_name": "your-project-name",
        "api_key": os.environ["ILLINOIS_CHAT_API_KEY"],
        "openai_key": os.environ["OPENAI_API_KEY"],
    },
)