Your Writing Software Stack
Table of Contents
Overview
In this course, you will have the opportunity to learn and use a set of professional tools that work together to support academic writing. Similar to how software engineers use a specific combination of tools (like a code editor, version control system, and testing environment), we’ll use a “stack” of writing software that helps with research, formatting, citation, and even idea generation. These tools are not just for this class. They are the kinds of software that many graduate students and researchers use every day in real academic and professional work.
Each tool plays a different role in the writing process. You don’t need to be an expert right away, and we’ll be incorporating each tool step by step as we need it in the projects we pursue. Think of this as part of your academic toolkit, helping you not just to write papers, but to manage sources, stay organized, and become more confident in your writing process.
Note that this guide is written for students at the University of Pittsburgh. It may contain useful information for those in other contexts, but instructors elsewhere may need to edit these materials to provide institution-specific guidance.
Hardware requirements
This course assumes you have access to a computer with a desktop operating system such as Windows or macOS. If you do not have your own computer, such as if you only have an iPad or other mobile device, please see: Virtual Student Computing Lab
Writing software tools you’ll be using
Library Search Tools: This semester, you will learn how to search effectively for academic sources using the university’s library catalog, PittCat, and other research databases.
Zotero: A free tool for collecting and organizing sources, generating citations, and creating reference lists automatically.
Microsoft Word: The main tool for writing and formatting your papers. We’ll learn how to use its built-in features like templates and styles to save time and meet academic formatting standards. See the Microsoft Word and PowerPoint Setup Guide. Special note: Instead of Microsoft Word, some students strongly prefer to use a LaTeX editor such as Overleaf. If you choose to use LaTeX in this course, keep in mind that assignment instructions are based on Microsoft Word and that I cannot provide support for LaTeX issues.
PowerPoint: Used for creating and practicing presentations, especially with tools like Speaker Coach that give feedback on your delivery.
Language Model (LLM): Tools like Claude, Gemini, or Copilot can help with brainstorming, outlining, and improving your drafts when used thoughtfully and ethically.
Check-In
Before we move forward in the course, it’s important to make sure your writing tools are set up and ready.
Click on the link to our university’s library catalog, PittCat. Bookmark the website so that you can find it again easily in the future.
Then, follow the guides to install Microsoft Word, PowerPoint, and Zotero. (You don’t need to install any local AI tools unless you’re curious.)
Then, after you have those programs installed, reply to the discussion board for this assignment with 2-3 sentences:
- Share one goal you have for your writing this term or one way you hope these tools will help you.
- Optionally, post a quick question if you’re unsure about anything in the setup process.
This activity is worth 2 points (complete/incomplete). For this first check-in, we want to hear your own personal goal in your own words. The purpose is for you to reflect on your needs and for us to get to know you better. Please no AI generated text in your response; only your own words.
Setup Guides
Microsoft Word and PowerPoint
Follow these steps to install the Microsoft Office suite, including Word, using your Pitt account:
On the Microsoft sign-in screen, enter your full Pitt email address, including
@pitt.edu, then clickNext.You’ll be redirected to the Pitt login page. Sign in using your Pitt username and password.
After logging in, you’ll be returned to the Microsoft 365 account page.
Look for the
Office apps & devicessection. Click theInstall Officebutton.Run the installer program once it downloads, and follow the prompts to complete installation.
When installation finishes, open Word. Sign in again with your Pitt email to activate Office.
You only need to install the apps you’ll actually use. Word and PowerPoint are the minimum requirements for this course. Other programs (Excel, etc.) are optional.
Zotero
Zotero is free software that helps you collect, organize, and cite sources for your academic writing. In this course, you’ll use Zotero to save the articles and resources you find, insert in-text citations into your Word documents, and automatically generate a properly formatted reference list in APA style. Once it’s set up, Zotero makes it easier to manage your sources and avoid citation mistakes, especially when working on longer writing projects with many references.
Use the steps below to install Zotero and get the Word citation plugin working:
Download Zotero. Go to https://www.zotero.org/download and choose the version for your computer (Windows, macOS, or Linux).
Install Zotero
- Windows/Linux: Run the installer and follow the prompts
- Mac: Open the downloaded file and drag Zotero into your Applications folder
Close Microsoft Word if it is open during installation.
Launch Zotero. The Microsoft Word plugin should install automatically the first time you open Zotero.
Open Word. You should see a
Zoterotab in the toolbar. This means the plugin is working.Recommended: Set up a Zotero account to enable Syncing. See the Zotero help page: Syncing
Troubleshooting
If the Zotero tab does not appear in Word:
- Open Zotero
- Go to Preferences
- On Windows/Linux:
Edit > Preferences - On Mac:
Zotero > Preferences
- On Windows/Linux:
- Select the
Citetab - Click
Word Processors, then clickInstall Microsoft Word Add-in - Restart Word and check again
Generative AI Tools
This resource is designed to help you identify and access several of the most widely used and relevant Large Language Model-based AI tools (LLMs) available. Whether you’re new to these tools or looking to explore beyond ChatGPT, this guide will provide you with the essentials: what each model is, how to access it, and how to start using it for academic and professional ideation tasks.
This guide includes four categories of models:
Tools Provided by Generative AI @ Pitt (university-supported tools with secure access for institutional use)
Other Frontier Models (freely accessible, web-based models not administered by the university)
Local Models on Desktop/Laptop (runs on your own device for complete privacy and offline use)
Local Models on Mobile Devices (a still-emerging format similar to, but generally less capable than, the local models on desktop/laptop)
1. Tools Provided by Generative AI @ Pitt
The University of Pittsburgh provides access to several generative AI platforms through Pitt Digital (See: Generative AI @ Pitt). These tools are vetted for institutional use and comply with university data protection policies. All are available for free and can be used with your Pitt credentials.
Claude (Anthropic) (🇺🇸)
Access: https://claude.ai using your Pitt email address
- Pitt students and faculty have access to all the “Claude for Education” enterprise features, including tools such as projects, artifacts, memory, and file uploads
- Requires an internet connection to use, but has a first-party desktop application: https://claude.ai/download
Microsoft 365 Copilot Chat (🇺🇸)
Access: https://m365.cloud.microsoft/chat/ using your Pitt email address
- In addition to the chat interface, Copilot tools are integrated directly into many Microsoft products such as Word and PowerPoint
- Chat interface requires an internet connection to use, but can be installed as a local web application: see Installing a web-based AI tool as a desktop app
Google Gemini (🇺🇸)
Access: https://gemini.google.com/app using your Pitt email address
- Conversational AI model integrated with Google Workspace
- Can generate text, assist with brainstorming, and interact with Google Drive files (Docs, Sheets, Slides, etc.)
- Requires an internet connection to use, but can be installed as a local web application: see Installing a web-based AI tool as a desktop app
Google NotebookLM (🇺🇸)
Access: https://notebooklm.google.com using your Pitt email address
- An AI research assistant; an interface where you can upload your own documents (such as research papers) to do document-based tasks, such as summarizing and answering questions
- Organizes sources and citations in a structured interface
- Requires an internet connection to use, but can be installed as a local web application: see Installing a web-based AI tool as a desktop app
All four of these tools are secured for appropriate use with Pitt-affiliated content. If you’re working with sensitive, institutional, research, or FERPA-protected data, you should prioritize these over third-party tools not governed by the university.
2. Other Frontier Models
This section lists some additional state-of-the-art (“frontier”) AI tools. These tools are not officially supported by Pitt Digital, but they are broadly used and accessible via free or paid accounts. They can be accessed through your web browser and require no installation.
ChatGPT (OpenAI) (🇺🇸)
Website: https://chatgpt.com
- Create a free account using your email address.
- Requires an internet connection to use, but has a first-party desktop application: https://chatgpt.com/download
DeepSeek Chat (DeepSeek) (🇨🇳)
Website: https://chat.deepseek.com
- Create a free account using your email address.
- Requires an internet connection to use, but can be installed as a local web application: see Installing a web-based AI tool as a desktop app
Kimi AI (Moonshot AI) (🇨🇳)
Website: https://www.kimi.com/
- Create a free account using your email address.
- Requires an internet connection to use, but can be installed as a local web application: see Installing a web-based AI tool as a desktop app
3. Local Models on Desktop/Laptop with LM Studio
While frontier models like ChatGPT, Claude, and DeepSeek run entirely in the cloud, it’s also possible to run large language models locally on your own computer. This gives you more control over your data and allows for offline use. The easiest way to get started is with a tool called LM Studio.
What is LM Studio?
Website: https://lmstudio.ai
LM Studio is a free desktop application that lets you download and run open-source LLMs on your personal device. It’s available for Windows, macOS, and Linux.
Features include:
- Easy installation of free, open-weight AI models.
- A simple built-in chat interface.
- GPU acceleration (if supported by your hardware).
- Ability to run fully offline, with no data sent to external servers.
- No sign-up required.
LM Studio is a great entry point for exploring edge models and understanding how modern AI works under the hood.
Recommended Model Families for Local Use
Several organizations have released high-quality open-weight models suitable for local deployment. Here are some notable families to explore as of late 2025:
- Gemma (Google, 🇺🇸): A family of efficient models designed for consumer hardware. Search for “gemma” in LM Studio’s Discover tab.
- Qwen (Alibaba, 🇨🇳): Known for strong performance across multiple languages. Search for “qwen” in LM Studio.
- Ministral (Mistral, 🇫🇷): Compact models from Mistral AI optimized for edge deployment. Search for “ministral” in LM Studio.
- Llama (Meta, 🇺🇸): One of the most widely-used open model families, with versions ranging from small to very large. Search for “llama” in LM Studio.
- Ernie (Baidu, 🇨🇳): Baidu’s contribution to the open-weight ecosystem, with particular strength in Chinese language tasks. Search for “ernie” in LM Studio.
- Phi (Microsoft, 🇺🇸): Small but capable models designed to run efficiently on limited hardware. Search for “phi” in LM Studio.
When browsing models in LM Studio, look for versions labeled with “GGUF” format and “Q4_K_M” quantization as a good starting point for balancing quality and performance.
System Requirements
The main system constraint for running an LLM is memory. I recommend using a system with at least 16 GB of RAM, preferably as video memory. LM Studio will notify you if a model is too large for your hardware.
Speeds depend heavily on factors such as GPU compatibility and memory architecture. Systems with discrete GPUs and systems with Apple Silicon chipsets tend to perform very well; but typical consumer PC laptops may perform more slowly.
Why Use Local Models?
Running models locally provides:
- Absolute privacy and security: All inputs and outputs stay on your device; no one else can store or spy on your data.
- Customization: You can experiment with different models and settings.
- Offline access: No internet connection is needed once the model is downloaded.
While local models may be slower or less capable than frontier models, they offer valuable security and flexibility and can be a useful part of your academic and professional toolkit.
4. Local Models on Mobile with PocketPal AI
As of 2026, it has become increasingly possible and practical to run local Language Models on mobile devices such as smartphones and tablets. However, in general, only quite small models (up to perhaps 2 billion active parameters) will be usable within the memory constraints of a typical mobile device, and acceptable speed and performance will be limited to newer or higher-end devices.
One popular chat interface for running local Language Models on mobile devices is provided by PocketPal AI, which is an open-source project which publishes apps for both iOS and Android. Similar to LM Studio mentioned in section 3 above, PocketPal AI is not itself a language model or chatbot. Rather, it implements the open-source llama.cpp runtime and provides a chat interface through which to interact with language models that the user downloads.
Inside of PocketPal AI, users can search for appropriate AI models to run. On an iPhone 17 Pro, I found that the model gemma-4-E2B-it-Q4_K_M (in GGUF form, as packaged by Unsloth) performed very well and produced acceptable output for general purposes.
Given the memory constraints of mobile devices and the limited sophistication of small Language Models, I would not consider this workflow to be ideal for professional or academic work. However, although this option for local AI inference pushes the limits of what the hardware and software ecosystem can currently provide as of 2026, it is plausible that this will soon be a robust, mainstream option for AI usage. A typical mobile device of model year 2030 may plausibly come pre-configured with local, private AI models that approximate the abilities of 2026’s frontier models such as ChatGPT, Claude, and Gemini, usable without any internet connection and running solely on local hardware.
Installing a web-based AI tool as a desktop app
Many of the most powerful and up-to-date Large Language Models (LLMs), like Google Gemini and DeepSeek Chat, do not offer a desktop application and are accessed through a web browser. While this makes them universally accessible, it can also mean your AI assistant gets lost in a sea of open tabs.
A better experience can be achieved using a dedicated, standalone application that you can launch directly from your Windows Taskbar or macOS Dock.
You can set up your own “desktop app experience” using a built-in browser feature called a Progressive Web App (PWA). A PWA essentially “wraps” a website in its own window and gives it a dedicated icon, providing an app-like experience for a web-only tool.
This guide will walk you through setting up Google Gemini as a PWA, but the same steps can be used for most other web-based tools.
How to Pin an AI App on Windows
For Windows 10 and 11, the most straightforward method is to use the built-in Microsoft Edge browser.
- Open the AI tool in Edge: Navigate to the AI tool (for example,
https://gemini.google.com/app) in Microsoft Edge and sign in if necessary. - Open the Menu: Click the three-dot menu (…) in the top-right corner of the browser window.
- Find the Install Option: In the dropdown menu, go to Apps and then select Install this site as an app.
- Confirm Installation: A small dialog box will appear. You can customize the name if you wish. Click Install.
- Pin to Taskbar: The PWA for the AI tool will immediately launch in its own window, and its icon will appear on your Taskbar. To make it permanent, right-click the icon on the Taskbar and select Pin to taskbar.
You can now close the window and launch the AI tool directly from its new icon at any time.
How to Add an AI App to Your Dock on macOS
For macOS, the recommended browser for this task is Safari, which offers seamless integration with the Dock.
- Open the AI tool in Safari: Navigate to the AI tool (for example,
https://gemini.google.com/app) in Safari and sign in if necessary. - Find “Add to Dock”: Go to the
Filemenu in the top menu bar (next to the Apple logo). - Select “Add to Dock…”: Click this option from the dropdown menu.
- Confirm: A dialog will appear, allowing you to customize the name. Click Add.
The icon for the AI tool will now appear in your Dock. Clicking it will launch the site in a clean, app-like window without the standard Safari browser interface.
Notes on Other Browsers
Google Chrome: Chrome has excellent PWA support on both Windows and macOS. The process is very similar to Edge’s. Simply navigate to the website and look for an install icon (a screen with a downward arrow) on the right side of the address bar, or find the “Install…” option in the main Chrome menu.
Mozilla Firefox: Firefox does not have a native, built-in feature for installing PWAs. There is a more limited feature called “Taskbar Tabs”, but this feature is available only for the Windows version of FireFox. Firefox users are advised to use one of the other browsers (Edge, Safari, or Chrome) for this specific task, or else explore third-party add-ins or extensions for Firefox that provide unofficial PWA support.
Virtual Student Computing Lab
When to Use This Tool
Many of our assignments will require you to use the full desktop versions of software like Microsoft Word and Zotero. If you are primarily using a device such as a tablet and do not have a personal laptop or desktop computer, the Virtual Student Computing Lab is the perfect solution for you. It gives you free, 24/7 access to a full Pitt computer from anywhere with an internet connection.
What It Is
The Virtual Student Computing Lab is a powerful tool that lets you connect to a university lab computer through your web browser or a simple app. It’s like having access to a computer in Hillman Library, but from your own device, wherever you are. You can use all the standard university software, install tools like Zotero for your own use, and save your work to your cloud storage (like OneDrive).
How to Access the Virtual Lab
Recommended Method: Use Windows App to Access the Virtual Student Computing Lab
- Download, install, and open “Windows App” from your platform’s store. See links below:
- Sign in with your Pitt email address by hitting the
+button and thenWork or School Account. You may be prompted with the standard Pitt Passport login after entering your Pitt email address. - Tap the
Pitt Digital - Virtual Student Computing Labbutton. - Enter your University Computing Account username and password to complete the login process.
Other method: Web access
You can connect to the Virtual Lab at any time by going to this simple, official Pitt URL:
You will be asked to sign in with your Pitt email address and password.
For More Information
Pitt Digital has a detailed guide with instructions for connecting from different devices (Windows, Mac, iPad, etc.) and answers to common questions. For a more detailed walkthrough, please see the official knowledge base article:
🔗 Virtual Student Computing Lab - Official Pitt Guide
Alternatives Worth Considering
The software stack described above represents a minimum viable setup that is widely used across academic contexts. However, different disciplines and individual preferences may favor alternative tools. This section briefly surveys other options worth knowing about, while noting that subsequent modules in this OER will focus specifically on Microsoft Word and Zotero workflows.
Alternative Desktop Word Processors
LibreOffice Writer
LibreOffice is a free, open-source office suite that includes Writer, a word processor fully compatible with Microsoft Word’s .docx format. LibreOffice works with Zotero’s citation plugin and provides similar formatting capabilities to Word. It runs on Windows, macOS, and Linux. For users who prefer open-source software or need a no-cost alternative to Microsoft’s subscription model, LibreOffice is a strong choice.
OnlyOffice
OnlyOffice is another free word processor with strong compatibility with Microsoft Word formats. It offers a clean interface and works with Zotero for citation management. OnlyOffice is available as desktop software and as a web-based collaborative platform.
Browser-Based Word Processors
Both Microsoft Word Online (accessible through Office 365) and Google Docs provide browser-based word processing with real-time collaboration features. These tools are convenient for working across devices and for group projects. However, citation management workflows can be more complex in browser environments. Zotero offers a browser connector for saving sources, but the citation insertion process may be difficult or impossible on web or mobile versions of Word. Google Docs users may need to use the Zotero Google Docs plugin or export to Word for final citation formatting.
LaTeX Workflows
For technical fields including mathematics, computer science, and physics, many researchers prefer LaTeX, a markup-based typesetting system that excels at handling complex equations, technical diagrams, and precise formatting. Overleaf is a popular web-based LaTeX editor that enables real-time collaboration. For citation management in LaTeX, BibTeX and BibLaTeX are the standard tools, offering powerful but syntax-intensive bibliography management. While LaTeX provides exceptional control over document formatting, it has a steeper learning curve than Word and may not be appropriate for all writing contexts.
Alternative Citation Managers
Mendeley
Mendeley is a widely-used citation manager with social networking features for researchers. It offers similar functionality to Zotero, including PDF management and Word integration. Mendeley is owned by Elsevier and is free for individual use, though not open source. In my own experience, Mendeley’s APA 7th Edition output has contained considerably more errors and issues than Zotero’ output, which is one reason I prefer the latter program.
EndNote
EndNote is a commercial citation manager popular in some research fields, particularly the sciences. It offers robust features but requires a paid license. Some universities provide institutional access, but individual licenses can be expensive. Zotero provides comparable functionality at no cost.
Closing Note
Any tool combination that supports your writing process effectively is valid. This OER focuses on Microsoft Word and Zotero because they represent the most common academic writing stack, are freely available to most university students, and offer strong accessibility features and cross-platform compatibility. Instructors adapting these materials for other contexts should feel free to modify tool recommendations based on disciplinary norms and institutional resources in your program.
Instructor Notes
Implementation guidance:
- Note that this guide is written for students at the University of Pittsburgh. It may contain useful information for those in other contexts, but instructors may need to edit these materials to provide institution-specific guidance.
Prerequisite knowledge:
- tips here
License: This chapter is licensed under CC BY 4.0.
Cite this chapter as: Price, B. (2026). Your Writing Software Stack. In Authorship & AI: Modular OER for Responsible Academic Writing with Generative Tools. University of Pittsburgh. https://billcprice3.github.io/authorship-and-ai/010-your-writing-software-stack.html