Skip to main content

AI Sandbox Pilot Project

In Summer 2025, OU Libraries was awarded an Artificial Intelligence Seed Grant from the Data Institute for Societal Challenges (DISC). Through this initiative, the Digital Scholarship and Data Services (DSDS) team is launching the OU Libraries AI Sandbox — a pilot ecosystem designed to give OU students, faculty, and researchers hands-on access to cutting-edge AI tools and infrastructure.

What We Offer

The AI Sandbox supports both no-code and API-based platforms, enabling you to learn, experiment with, and integrate AI into your research, teaching, and creative work. You'll have hands-on access to a range of generative and non-generative AI tools, including LibreChat, LiteLLM, HuggingFace, AWS Bedrock, and National Research Platform (NRP). We provide training materials, workshops, and three high-performance workstations to support your learning and experimentation with these AI technologies.

Get Started

Who Can Use This Service

This service is open to all OU students, faculty, and researchers.

What is Included

The AI Sandbox Pilot Project includes five core services:

1. LibreChat (No-Code ChatGPT-Style Interface)

LibreChat provides an accessible chat interface to interact with both open-source and commercial large language models. It’s designed for researchers, instructors, and students who want to explore AI capabilities without writing code.

Features include:

  • Clean, ChatGPT-style UI
  • Model switching between National Research Platform (NPR) and AWS Bedrock models
  • Multimodal support (text and/or image inputs)
  • Persistent session history and export chat history
  • Secure OU login via the university’s SSO credentials

Use Cases: brainstorming, drafting, research ideation, multimodal interaction, workshops, and classroom instruction.

Visit our LibreChat Guide for documentation on how to use this technology.

For information about our terms of service and privacy policy, please visit our LibreChat Policy page.

Log On To LibreChat

2. LiteLLM – API Access for Advanced Users

LiteLLM provides OpenAI-compatible API access to multiple AI models, enabling programmatic workflows for research, development, and automation.

Features include:

  • Access to AWS Bedrock and NRP-hosted models
  • Customizable to create teams and to use virtual API keys
  • Ability to set cost limits and track usage for admins
  • Multimodal support (text, images, file URLs, code snippets)
  • Can run resulting output on other platforms or computer systems

Visit our LiteLLM Guide for documentation on how to use this technology.

Explore LiteLLM

3. Hugging Face + Nautilus – Open-Source AI at Small-Scale

Leveraging OU Libraries’ Nautilus Platform and HuggingFace, this service provides scalable, containerized environments for training, and running open-source AI models.

Features include:

  • JupyterLab Python environment with PyTorch, TensorFlow, and Hugging Face
  • GPU-enabled workflows through the National Research Platform (NRP)
  • Support for both exploratory and production research projects
  • Secure access for OU-affiliated users

Visit our Nautilus + HuggingFace Guide for documentation on how to use this technology.

4. AI Workstations – Local Research Infrastructure

OU Libraries has deployed three high-performance workstations (Mac Studio and Dell Pro Tower) equipped with advanced GPUs to support AI research and education.

Key Specifications:

  • NVIDIA RTX 5000 ADA GPU
  • 2 * M3 Ultra Mac Studio with 60-core GPU

We will provide workstation locations and booking details as the project develops!

5. AI Workshops and Carpentries

The Libraries offer both foundational and advanced workshops to help the OU community effectively use the AI sandbox technologies. Workshops are built on existing Carpentries programs and customized AI sessions.

Topics:

  • Intro to Unix Shell & Python
  • Programming with LLMs (APIs and Python)
  • Docker & Containers for AI
  • Responsible AI and Academic Integrity

All workshops are listed on the OU Libraries Events Page, with registration links and session descriptions.

Contact Information

If you have questions, need assistance, or would like to collaborate, please reach out to us:

Digital Scholarship and Data Services Team

Email: dsds@ou.edu

AI Sandbox Feedback Form