Introduction to Academic Support Services for Local AI
Modern research often involves datasets too large and complex for manual analysis. The Baylor University Libraries' Academic Support Services for Local AI is designed to assist with this challenge by leveraging the power of generative AI directly to your research data and documents.
This service uses local AI, which means all analysis is performed on your personal computer or on secure university hardware, not on public cloud services. This approach keeps your sensitive research data private and secure, while avoiding costly cloud computing fees.
Key capabilities include:
- Automate the coding and thematic analysis
Of qualitative data from interview transcripts or open-ended survey responses. - Summarize, classify, or identify key topics
Across large document collections, from research articles to historical archives. - Enrich, classify, or reformat data
In large spreadsheets by applying custom, repeatable instructions to every row.
Benefits
By running AI models on local hardware, this service offers a unique combination of advantages for your research workflow.
- Scale and Speed
This service allows you to analyze large datasets much more quickly than is possible by hand, which can support new research questions. - Data Security and Privacy
Because your files are never sent to a third-party, you can use these tools on sensitive or confidential data while maintaining privacy. - Cost-Effectiveness
The service is available at no cost to you, which avoids the service fees typically required by commercial AI platforms. - Customization and Control
You have the flexibility to create custom workflows, helping to ensure that your research methods are transparent and easily reproduced. - Full Support
You are not on your own. The Library’s Data & Digital Scholarship unit provides full support for this service, from initial consultation to troubleshooting.
Getting Started
The best way to begin is with a simple conversation. This is a fully supported service, and we recommend sending a quick email to dig_schol@baylor.edu to discuss your project. We can help you select the right path and tools from the very beginning. When you're ready to dive in, follow these steps to get your project up and running.
Step 1: Choose Your Environment
Your research needs will determine the best place to conduct your work. We support three primary environments:
- Your Personal Computer
For getting started, prototyping, and running smaller-scale analyses on your own Windows or Mac computer equipped with a modern GPU. - The Design & Data Lab
Our primary lab offers 16 powerful workstations and is the best place for hands-on consultations, workshops, and group projects. Please be aware that this space is sometimes reserved for instruction. - The Design & Data Lab II (24/7 Access)
For uninterrupted, round-the-clock access, this lab provides four additional workstations. It is located in the 24/7 space of the library and is never reserved, making it a reliable option for after-hours work or when the main lab is in use. - Kodiak High-Performance Computing Cluster
For large-scale, demanding projects that require the analysis of thousands of documents or massive datasets.
Step 2: Select Your Software
The ecosystem of local AI software is constantly growing. To provide the best support, we have focused our expertise on two powerful and reliable tools. They are designed for different types of users and workflows.
For a Graphical Interface: Use Jan.ai
Jan.ai provides a user-friendly, point-and-click application. We recommend this tool for anyone who is new to using local AI or for researchers who want a simple way to prototype and test different models and instructions before building a larger, automated workflow.
For Automation and Scripting: Use Ollama
Ollama is a powerful command-line tool designed for users comfortable working in a terminal. It is the ideal choice for integrating AI into scripts for large-scale, repeatable research tasks. We have developed many numerous starter scripts to help researchers get up and running in no time. This makes it the fastest path for getting our pre-configured workflows up and running.
Step 3: Run Your First Analysis
To help you get started as quickly as possible, we have created six pre-configured scripts. These scripts are ready to use and are tailored for running analyses on personal computers (Windows/Mac) or on the Kodiak cluster. We will provide you with these scripts and instructions during your initial consultation.
What Type of Help is Available
The Library’s Data & Digital Scholarship unit provides full support for this service. We are available to partner with you on your research and teaching in several key ways.
- One-on-One Project Consultations
We can help you brainstorm ideas, choose the right tools and models for your specific data, and adapt our pre-built scripts for your project. - Hands-On Prototyping in the Lab
Use the powerful workstations in the Design & Data Labs to test your workflows with our direct support and guidance. - Instruction and Classroom Support
We can partner with you to bring these tools into your classroom through guest lectures, custom workshops for your students, or by helping you develop assignments. - Custom Script Assistance
For unique research challenges, we can provide guidance and support to assist developing custom scripts tailored to your specific needs.
