OwlChat (Open WebUI)

What do you need help with?

About FAU OwlChat Learn about OwlChat, OpenWebUI capabilities, and available models

Getting Started & Access How to log in via SSO and launch your first AI conversation

Features & OpenWebUI Tools Document uploads (RAG), prompt shortcuts, and side-by-side chats

Using Gemma4-vibe Detailed guide and prompt templates for the featured model

Data Privacy & Guidelines Responsible AI usage, academic integrity, and sensitive data

Get HPC Support Contact FAU High Performance Computing & Help Desk

About FAU OwlChat

Learn about OwlChat, the OpenWebUI interface, and available models.

FAU OwlChat (https://chat.hpc.fau.edu) is Florida Atlantic University’s web-based artificial intelligence portal provided by the High Performance Computing (HPC) team. Powered by OpenWebUI, OwlChat delivers an intuitive, feature-rich interface for research, coursework, software development, and administrative productivity.

OpenWebUI is an open-source, user-friendly WebUI that acts as the frontend interface for running AI language models. Key highlights include:

ChatGPT-Style Interface: Includes thread management, Markdown formatting, syntax highlighting for code, and customizable user profiles.
Document Retrieval (RAG): Attach files directly into your prompt to analyze, summarize, or query specific documents.
Model Comparison: Run two or more models in parallel to evaluate differences in response structure and quality.
Privacy First: Self-hosted within FAU infrastructure to maintain academic and institutional privacy.

OwlChat offers a flexible model ecosystem depending on computational availability and access permissions:

Featured Local Model (Gemma4-vibe): Optimized for high throughput, creative drafting, research summaries, and interactive coding. It serves as the recommended default for most users.
Additional Cloud Models: Many cloud models are also available via the top selection menu for users who require specialized domain models or larger parameter configurations.

Getting Started & Access

How to sign in and begin using OwlChat with your FAU credentials.

Access to OwlChat is integrated with FAU Single Sign-On (SSO) and Multi-Factor Authentication.

Before connecting to OwlChat, verify that you have:

An active FAUNet ID and password.
Access to your registered Microsoft Multi-Factor Authentication (MFA) device.
A modern web browser (Google Chrome, Mozilla Firefox, Apple Safari, or Microsoft Edge).

Go to https://chat.hpc.fau.edu.
Click Sign in with Single Sign-On (SSO).
Enter your full FAU email address (e.g., FAUNetID@fau.edu) and click Next.
Enter your FAU password on the central authentication page.
Approve the Microsoft MFA prompt on your mobile authenticator app.
Upon successful login, you will land on the OpenWebUI chat workspace.

Note: If you receive an authorization error during initial sign-in, your HPC group permissions may need to be updated. Submit a ticket via the FAU Help Desk portal.

Features & Interface Guide

Learn how to leverage OpenWebUI features for maximum productivity.

OpenWebUI provides advanced tools designed specifically for technical and academic workflows.

Top Bar / Model Selector: Choose your active model (such as Gemma4-vibe) or select multiple models to perform parallel outputs.
Left Sidebar: View past chat histories, rename sessions, group chats by tags, or start a New Chat.
Prompt Field: Type prompts, use shortcut slash-commands (/), or drag and drop files directly into the window.

OwlChat supports Retrieval-Augmented Generation (RAG), enabling models to cite and query uploaded files directly.

Click the + (Attachment) icon next to the prompt bar or upload documents via the Workspace -> Documents section.
Upload supported file formats such as PDF, TXT, MD, CSV, or DOCX.
Type # followed by the document name in your chat prompt to reference that specific document directly in your discussion.

Save repetitive system prompts or guidelines for quick re-use:

Navigate to Workspace -> Prompts to create custom system commands.
Assign a slash-command trigger (e.g., /summarize or /code-review).
In any active conversation, type / to quickly inject your customized system instructions into the prompt.

Using Gemma4-vibe (Featured Model Example)

Practical walkthrough and prompt templates using the Gemma4-vibe model.

Gemma4-vibe is the primary featured model on OwlChat, tuned for conversational speed, technical synthesis, and high-accuracy task execution.

Click New Chat in the upper-left navigation panel.
Open the model selection dropdown menu at the top of the chat area.
Select Gemma4-vibe from the available model list.
Verify that Gemma4-vibe appears as the active model tag above the prompt field.

Copy and adapt these sample prompts when working with Gemma4-vibe:

1. Python Code Development & Error Analysis:

"Write a Python script to parse an HPC job log file, extract all memory error occurrences, and calculate the total execution runtime. Add clear code comments."

2. Academic Literature Review & Synthesis:

"Summarize the uploaded research PDF in three sections: Background, Core Findings, and Future Work. Present the summary using structured bullet points."

3. Concept Simplification & Study Help:

"Explain the difference between message passing interface (MPI) and shared-memory multithreading (OpenMP) as if I am an undergraduate computer science student."

Data Privacy & Usage Guidelines

Important guidelines on data security and academic integrity.

Protected Health & Student Data: Do not submit HIPAA-regulated health data, Social Security numbers, or FERPA-protected confidential records.
Academic Integrity: Students must follow course-specific AI guidelines established by their instructors before using OwlChat for graded coursework.
Fact Verification: AI models may generate inaccurate information or hallucinations. Always verify critical facts, references, and mathematical equations.

Updating...