Choose the model
Pick the model that suits the job — Claude, ChatGPT or Gemini — and switch at any time. Each has its own strengths, so an agent built for long-form writing feedback need not use the same model as one built for data questions.
Ground it in your material
Upload the sources your agent should rely on: PDFs, Word documents, PowerPoint decks, spreadsheets and plain text, or connect Google Drive, OneDrive, Dropbox, a website, a GitHub repository or a video with captions. Clementine reads text, so text-based sources give the best results — a scanned PDF or an image-heavy slide has little for it to work with.
Set the tone and the guardrails
System prompts define the agent's purpose and voice. Proactive and guided prompts give people an obvious place to start. Moderation and safety settings let you filter inappropriate input and keep the agent from wandering into medical, legal or financial advice — redirecting to the right campus resource instead.
Control who sees it
Decide whether an agent is visible to students, to administrators or to anyone, and whether it requires a Syracuse sign-in to chat. When it is ready for a wider audience, you can embed it directly in a webpage.
Build on it programmatically
Clementine has an API for working with agents and their datasets in your own code. To request access, email aihelp@syr.edu from your Syracuse account with your NetID and a short description of what you plan to build.