# SU ITS Data & AI Knowledge Base - Full Markdown Corpus This file concatenates the public Markdown pages for agents that need a single context source. For normal question answering, fetch `/llms.txt` first and then fetch the one or two most relevant `.md` page URLs. --- title: "Approved Tools for Use with University Data" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/approved-tools-for-use-with-university-data.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/approved-tools-for-use-with-university-data.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488144948/Approved+Tools+for+Use+with+University+Data" page_id: "488144948" tags: ["policy", "overview", "data-privacy", "enterprise", "access"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2025-10-27" --- When a user is logged in with Syracuse University credentials (NetID), Syracuse University supports the use of the following University approved tools for university data: - [Anthropic Claude Enterprise](https://claude.ai/) - Request Access at [getclaude.syr.edu](https://getclaude.syr.edu) - [mentorAI for Syracuse University](https://mentor.ai.syr.edu/) - [Microsoft Copilot Work or Web](https://www.office.com/?auth=2) - [Google Gemini](https://gemini.google.com/app) - login with NetId@g.syr.edu - [Blackboards AI](https://blackboard.syracuse.edu/) - [OpenAI ChatGPT Teams](https://chatgpt.com/) - Request access via IT Support - [Adobe Firefly](https://firefly.adobe.com/) - Gradescope --- title: "Creative AI Workflows & Tools" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/creative-ai-workflows-tools.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/creative-ai-workflows-tools.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/515670055/Creative+AI+Workflows+&+Tools" page_id: "515670055" tags: ["example-use", "creative", "claude", "notebooklm", "copilot"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-30" --- Discover powerful ways to use AI tools beyond basic questions and answers. Each tool has unique capabilities - from building interactive dashboards to generating podcast-style summaries. Below are the superpowers of each tool and creative workflows that combine them in surprising ways. For full tool details and data policies: [Approved Tools for Use with University Data](./approved-tools-for-use-with-university-data.md) --- # Tool Superpowers: ### 🧠 Claude - The Builder > [!info] > Superpower: Builds interactive tools you can use and share - not just text responses > > Unique Capability: Creates live artifacts (calculators, dashboards, visualizations, quizzes) that you edit in real-time > > Best For: > > - Building interactive tools without coding > - Deep analysis of long documents (200+ pages) > - Technical work and code generation > > Try It: Ask Claude to "build an interactive grade calculator" or "create a filterable task tracker" > > Claude FAQ: [Claude - Frequently Asked Questions](./claude-frequently-asked-questions.md) --- ### πŸŽ™οΈ Gemini NotebookLM - The Podcast Machine > [!info] > Superpower: Turns documents into 10-20 minute AI-generated podcast conversations > > Unique Capability: Audio overviews - two AI hosts discuss your content in engaging dialogue format > > Best for: > > - Making dense content listenable (commute/gym study) > - Multi-document research with citations > - Creating audio study guides > > Try It: Upload lecture transcripts and generate an audio "course overview" --- ### βš™οΈ Microsoft Copilot for Work - The Auto-Documenter > [!info] > Superpower: Automatically documents Teams meetings while you focus on participating > > Unique Capability: Zero-effort meeting summaries - records, transcribes, and summarizes without manual work > > Best for: > > - Microsoft 365 integration (Word, Excel, PowerPoint, Teams) > - Enterprise-level automation > - Organization-wide search > > Where You'll Find Copilot Buttons: > > - Microsoft Teams - Meeting transcription, chat summaries, catch-up on missed meetings > - Word - Draft documents, rewrite sections, summarize long files > - Excel - Analyze data, create formulas, generate insights > - PowerPoint - Create presentations from prompts, design slides > - Outlook - Draft emails, summarize long threads > - OneNote - Organize notes, create summaries > > **⚠️ Important:** Microsoft Copilot for Work requires an M365 Copilot for Work license, which is an additional license that must be purchased. > > Try It: Let Copilot for Work watch your meeting, then ask "What were the key decisions?" --- # Creative AI Ideas --- > [!note]- The Audio Time Machine > Upload entire semester of lecture transcripts to NotebookLM > > Generate single course overview podcast (10-20 min) > > Listen before finals = comprehensive review on commute > > Turn 40 hours of lectures into digestible audio > [!note]- The Living Dashboard Builder > Upload meeting notes to Claude > > Ask: Build interactive task tracker with filters and progress bars > > Get working artifact you can share with team > > Update it by talking to Claude, not coding > > Project management without the software > [!note]- The Lecture-to-Flashcard Pipeline > Record class with permission > > Upload recording(s) to mentorAI mentor: Create 30 flashcards + 10 practice questions with answers > > Upload same transcript to NotebookLM for audio study guide > > Study with both active recall AND passive listening > [!note]- The Email Project > Ask Copilot for Work: Find all emails about [project] from last 6 months > > Ask Claude: Find all emails about [project] from last 6 months (Requires the enabling of the Microsoft 365 Connector) > > Create timeline of decisions made > > Draft status update based on this history > [!note]- The AI Debate Arena > Claude argues Position A, mentorAI-ChatGPT argues Position B > > You moderate by copying responses between them > > Watch AIs strengthen each other's arguments > > See both sides at their best, not their worst > [!note]- The Grade Calculator You'll Actually Use > Upload syllabus to Claude > > Ask: Build interactive grade calculator with 'what-if' sliders > > Get working tool: If I get 85% on final, what's my grade? > > Share link with classmates > > Bookmark and use all semester > [!note]- The Prompt Breeding Lab > Ask mentorAI: Critique this prompt and improve it > > Feed improved version to Claude: Make this even better > > Feed that to Gemini for third perspective > > Watch prompts evolve across AI ecosystems > > Learn what makes prompts actually work > [!note]- The Research Poster Generator > Upload paper to Claude: Create conference poster layout in HTML > > Get interactive artifact with your data visualized > > Edit colors and text in real-time > > Export and print = professional poster in 10 minutes > [!note]- The Zero-Effort Meeting Documentation > Copilot for Work auto-transcribes your Teams meeting > > Upload transcript to Claude: Use Meeting Transcript project > > Get shareable task tracker artifact > > From meeting to project board in 2 minutes > [!note]- The Personal Research Assistant > Create a custom AI assistant with your citation style > > Upload papers and have it summarize in your format > > Build a library of summaries in consistent style > [!note]- The Citation Style Switcher > Upload your paper to Claude > > Ask: Create tool that converts citations between APA/MLA/Chicago > > Get interactive converter artifact > > Paste citation, get all three formats > > Never look up citation rules again > [!note]- The Multi-Document Synthesizer > Upload 10 papers to NotebookLM > > Ask: What do these papers disagree about? Where's the gap? > > Get literature review with identified conflicts > > Find your research question from what's missing --- # Support Need help? ITS Help Desk: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Hermes Agent: What It Is and How to Approach It Safely" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/hermes-agent-what-it-is-and-how-to-approach-it-safely.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-general-information/hermes-agent-what-it-is-and-how-to-approach-it-safely.md" tags: ["ai-agents", "security", "data-privacy", "third-party-tools"] audience: ["students"] last_modified: "2026-06-08" --- Hermes Agent is a popular open-source AI agent that students may run into and want to experiment with. This page explains what it is, how it differs from assistants like Claude or Copilot, and the security realities you accept the moment you run it, so you can decide whether and how to try it without putting yourself or the university at risk. --- ## What is Hermes Agent Hermes Agent is an open-source autonomous agent built by Nous Research and released in February 2026 under the MIT license. Unlike a chat assistant you open in a browser, it runs as an always-on background process (a daemon) on a computer or server you control. It keeps memory across sessions, runs scheduled jobs on a timer (cron), connects to messaging apps such as Telegram, Discord, Slack, and WhatsApp, and can write its own reusable "skill" files so it gets more capable the longer it runs. It is provider-agnostic: you point it at a language-model backend (Anthropic, OpenAI, Google, OpenRouter, a Nous subscription, or a local model). It needs a Unix-like environment, so on Windows you have to install WSL2 and run it from inside that. The distinction that matters: Claude and Copilot are assistants you prompt one message at a time. Hermes is an agent you deploy. It has a terminal, file access, and the ability to act on a schedule without you watching. That capability is both the appeal and the entire risk. --- ## Why students are interested The draw is real. Hermes gives you persistent memory so you stop re-explaining context, automation that runs while you sleep, a single agent reachable from the messaging app you already use, and all of it free and self-hosted. For anyone studying AI agents, systems administration, or security, it is a legitimate hands-on learning platform, not a toy. The same features that make it useful are exactly the ones that make it dangerous if you run it carelessly, which is the rest of this page. --- ## The security model you are accepting > [!warning] You are running a program that can act on its own > Running Hermes means running software that can execute shell commands, read and write files, reach the internet, and take scheduled actions on your behalf, continuously. Treat the machine it runs on as fully exposed to it. ### It is permissive by default An independent security audit of an earlier release found no malware and no hidden data exfiltration, but it concluded that the out-of-the-box posture is allow-all, which is risky for anyone who does not know to harden it. Authorization is binary: an authorized user gets full terminal, file-write, and scheduling access, and there is no built-in tier that grants someone "read-only" or "chat-only" use. There are also per-platform flags that authorize every user at once, and a "YOLO" mode that auto-approves every command it wants to run. Set command approval to manual, and never turn on auto-approve on a machine you care about. ### The only real boundary is your user account Nous Research's own security policy is explicit that the load-bearing boundary is the host operating system account Hermes runs as: its file permissions and what that account is allowed to reach. The in-process checks (command pattern-matching, prompt-injection scanning of context files) are helpful, but the project does not treat them as guarantees. The practical consequence is that you should run Hermes as a low-privilege user in an isolated environment, a dedicated virtual machine, container, or WSL2 distribution, and never as an administrator or root account on your daily computer. Assume that anything that account can touch, the agent can touch. ### The agent writes its own skills, and that is an attack surface The audit flagged that skill files the agent creates persist and get loaded into future sessions, which makes them a durable prompt-injection vector, and that the guard screening those files is pattern-based and tends to "ask" rather than "block" on agent-written content. In plain terms, content the agent reads today (a web page, an email, a file you hand it) can plant instructions that fire later. Review any skills it writes, and be careful about pointing it at content you do not trust. ### Windows users silently lose a layer One of Hermes's command-screening tools (called tirith) does not run on native Windows and is skipped without showing you a warning. The documented fix is to run Hermes under WSL2, which it requires anyway. If you are on Windows, use WSL2, both because Hermes needs it and because skipping it quietly disables a safety check you would otherwise have. --- ## If you want to try it safely > [!note] Treat this as a checklist, not a suggestion > Hermes is only as safe as the environment you put it in and the limits you set. The defaults will not protect you. 1. **Use a throwaway, isolated host.** A dedicated virtual machine, a cheap cloud instance, a container, or a WSL2 distribution. Do not run it in your primary operating-system account, and never as administrator or root. 2. **Run as a low-privilege user.** The OS account is the real boundary, so give it the least access it needs to do what you are testing. 3. **Set approvals to manual and leave auto-approve off.** Review commands before they run, especially while you are still learning what it does. 4. **Set an allowlist.** Network-connected adapters should refuse to act until you tell them who is allowed. Lock messaging access to your own user ID only. 5. **Protect your credentials.** Keep the configuration and secrets file readable only by you (`chmod 600`), use a separate API key per service, and never commit secrets to a git repository. 6. **Never grant send or delete authority you would not give a stranger.** For email or messaging, configure it to draft actions for your review rather than executing them. 7. **Keep its data off synced folders.** Background file-sync services like OneDrive or Dropbox fight a daemon over file locks and can corrupt the agent's state. Store its data on the local filesystem. 8. **Back up its state, and treat that backup as a secret too.** It can contain tokens and personal context. --- ## Do not connect it to university accounts, data, or systems > [!warning] This is the hard line for Syracuse students > Hermes is not an approved tool for use with university data, and its allow-all, single-boundary design is the opposite of what handling protected information requires. Do not authenticate Hermes with your NetID or any Syracuse account. Do not give it access to university email, files, Teams, or any other SU system. Do not run it on university-managed or university-owned hardware. Keep your experimentation on personal accounts and personal hardware, fully separated from anything connected to Syracuse. If you have a legitimate academic or research reason to run an autonomous agent against any university resource, that decision goes through SU IT and the approved-tools process first, not the other way around. See [Approved Tools for Use with University Data](approved-tools-for-use-with-university-data.md). --- ## Keep in mind - **You are the security boundary.** Hermes is as safe as the environment you give it and the limits you set, not as safe as its defaults. - **Verify before you trust.** It can act on bad or injected instructions, so review what it does, especially anything it scheduled or any skill it wrote for itself. - **Personal and university stay separate.** Nothing connected to Syracuse should touch Hermes, and nothing Hermes touches should go near your Syracuse credentials. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 Β· [help@syr.edu](mailto:help@syr.edu) --- title: "AI @ Syracuse University" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-syracuse-university.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/ai-syracuse-university.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/483525103/AI+@+Syracuse+University" page_id: "483525103" tags: ["overview", "getting-started", "claude", "mentorai", "policy"] audience: ["students", "faculty", "staff"] last_modified: "2026-03-20" --- ## Welcome to the AI at Syracuse University Generative artificial intelligence (AI) tools like Claude by Anthropic, Microsoft Copilot and Google Gemini have redefined what it means to work, study, teach and create. These tools generate new content based on your prompts and if you bring it your data. #### Stay up to date on the changing AI landscape with the ITS [AI Insights Newsletter](https://its.syr.edu/news-events/ai-insights/) --- ## Tools for Students, Faculty and Staff 🧠 **Claude Enterprise** Claude Enterprise is an advanced AI assistant developed by Anthropic and is available to all Syracuse University students, faculty and staff. [Learn more about Claude](./claude-enterprise-at-syracuse-university.md) [Sign-up Now!](https://getclaude.syr.edu) ⭐ **Clementine Platform - mentorAI** Syracuse Universitys private AI platform for Teaching Learning and AI Innovation. Available to all Syracuse University students, faculty and staff. [Learn more about the Clementine Platform](./mentorai-syracuse-university.md) [Try Clementine](https://mentor.ai.syr.edu/platform/syracuse/e34f0313-81fc-4156-9cb1-aa7e0b90a3ce) ✨ **Microsoft Copilot** A versatile assistant built with OpenAI's latest GPT-5 model (the same model that powers ChatGPT). [Try Copilot](https://m365.cloud.microsoft/chat) [FAQ](https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488571112/Copilot+-+Frequently+Asked+Questions?atlOrigin=eyJpIjoiNTc2MzE0NTlmZjdjNGY1NmFmOTRjNmJmNzQ0MzhmMmYiLCJwIjoiYyJ9) πŸ’Ž **Google Gemini** An AI assistant developed by Google, Syracuse has access Google's 2.5 Pro Model which features guided learning, image generation, and a canvas editor. [Try Gemini](https://gemini.google.com/app) [FAQ](https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488865911/Gemini+-+Frequently+Asked+Questions?atlOrigin=eyJpIjoiMGVkMmM5NGMzMGI2NGI2NzgwOWYwZjI3NWZmZjQ5NWIiLCJwIjoiYyJ9) ### [Additional AI tools and platforms approved for use](./approved-tools-for-use-with-university-data.md) --- ### Ready to jump into AI, tired of reading? ![Screenshot 2026-03-10 203239.png](./attachments/483525103/Screenshot 2026-03-10 203239.png) Watch our Tech Tips video series on [Generative AI](https://video.syr.edu/playlist/details/1_spy966ue) --- ## Search this Knowledge Base --- ## Recently updated content --- title: "Claude Code Setup" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-code-setup.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-code-setup.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/986841103/Claude+Code+Setup" page_id: "986841103" tags: ["claude-code", "setup", "how-to", "getting-started", "access"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-04-15" --- # Step 1: Get access to Claude Code Using one of the methods for obtaining Claude Code described on the page: [Purchase Claude Code and Claude API Access - Artificial Intelligence (AI) - Answers](./purchase-claude-code-and-claude-api-access.md) # Step 2: Install Claude Code & Development Tools An automated PowerShell script for Windows (`Install-DevTools.ps1`) downloads and configures portable versions of all required development tools. The script auto-detects your system architecture (AMD64 or ARM64) and downloads the appropriate installers. It's idempotent, so it's safe to run multiple times. ### What Gets Installed | **Tool** | **Version** | **Notes** | | --- | --- | --- | | Node.js | 24.12.0 | Portable/embedded | | Git Portable | 2.52.0 | Includes Git Bash | | Python Embeddable | 3.14.3 | pip-enabled; skipped if system Python detected | | VS Code Portable | Latest | Skipped if already installed on the system | | GitHub Desktop | Latest | Visual interface for Git version control | | PowerShell 7 | Latest | Installed via winget from Microsoft Store; skipped if already installed | | Claude Code | Latest | Always installs locally; requires Git first | ### Where Tools Install By default, tools install to your **OneDrive** for automatic sync across devices. Claude Code always installs locally regardless of this setting. | **Location** | **Path** | | --- | --- | | OneDrive (default) | `%OneDrive%\Apps-SU\[arch]\` | | Local | `%USERPROFILE%\Apps-SU\[arch]\` | | Claude Code (always) | `%USERPROFILE%\.local\bin` | Tools are organized into architecture-specific subdirectories to prevent conflicts when OneDrive syncs across devices with different architectures: ``` Apps-SU\ β”œβ”€β”€ AMD64\ β”‚ β”œβ”€β”€ Node\ β”‚ β”œβ”€β”€ PortableGit\ β”‚ β”œβ”€β”€ Python\ β”‚ └── VSCode\ └── ARM64\ β”œβ”€β”€ Node\ β”œβ”€β”€ PortableGit\ β”œβ”€β”€ Python\ └── VSCode\ ``` ## Set Execution Policy Before running the script, open PowerShell and run: ```powershell Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser ``` ## Download & Run the Install Script Download the script, save it as `Install-DevTools.ps1` [Install-DevTools.ps1](./attachments/986841103/Install-DevTools.ps1) ## Remove the Mark of the Web When you download a file from the internet, Windows tags it with a "Mark of the Web" (MOTW) that blocks it from running, even if your execution policy is set to RemoteSigned. You need to unblock the file before running it. **Option 1: PowerShell command (recommended)** Open PowerShell and run: ```powershell Unblock-File -Path "$HOME\Downloads\Install-DevTools.ps1" ``` Adjust the path if you saved the file somewhere other than your Downloads folder. **Option 2: File Properties** 1. Right-click the downloaded `Install-DevTools.ps1` file in File Explorer 2. Select **Properties** 3. At the bottom of the General tab, check the **Unblock** checkbox 4. Click **OK** ### Run the Script Run it from PowerShell: ```powershell # Install all tools to OneDrive (default) .\Install-DevTools.ps1 ``` # Step 3: Configure Claude Code Once installation is complete, follow these steps to connect Claude Code to your Syracuse University account 1. **Launch Claude Code.** Open a terminal and type: ``` claude ``` 1. **Choose your text display style.** Select whichever option you prefer. 2. **Select "Anthropic Console account"** (option 2) when prompted for your login method. 3. **Sign in.** A browser window will open, and sign in with your `netid@syr.edu` credentials. If prompted, select the **Syracuse University** organization. 4. **Authorize Claude Code** when prompted in the browser. 5. **Return to the terminal** and press Enter twice to accept and trust the folder **You're all set!** Claude Code is now ready to use. Start by asking it a question or giving it a task in the terminal. # Optional: GitHub GitHub is an excellent companion to Claude Code. When Claude Code makes changes to your files, Git tracks every modification so you can see exactly what was added, removed, or changed. This is especially valuable when working with an AI coding tool because it gives you a clear record of what Claude did and makes it easy to undo any changes you don't want to keep. **Why use Git with Claude Code?** Git gives you a safety net when letting Claude Code work on your files. Before you ask Claude to make changes, you can commit your current work. After Claude finishes, you can review a line-by-line diff of everything it touched. If something doesn't look right, you can revert back to your previous commit in seconds. Without Git, you would have no easy way to see what changed or roll things back. Git also makes it practical to experiment. You can create a branch, let Claude Code try an approach, and if it doesn't work out, simply switch back to your main branch. This encourages you to use Claude Code more boldly, knowing you can always get back to a known good state. **Getting started with GitHub** [GitHub at Syracuse University - Information Technology Support - Answers](https://answers.atlassian.syr.edu/wiki/search?text=GitHub+at+Syracuse+University&spaceKey=ITHELP) # Optional: VS Code Integration For an enhanced development experience, you can integrate Claude Code with Visual Studio Code. Follow the official documentation: [Claude Code VS Code Integration Guide](https://code.claude.com/docs/en/vs-code) # Troubleshooting This problem typically arises when the script is downloaded from the web (as opposed to cloning the repo via Git). Windows applies a "Mark of the Web" to downloaded files, which blocks PowerShell from executing them. To fix this, unblock the file using one of the methods described in Step 2 above. If you prefer not to unblock the file, you can also bypass the execution policy for a single session: In PowerShell: `Set-ExecutionPolicy -ExecutionPolicy Bypass -Scope Process` From the Command Prompt (cmd): `powershell -ExecutionPolicy Bypass -File .\Install-DevTools.ps1` The script only supports AMD64 and ARM64. 32-bit systems are not supported. Restart your terminal or PowerShell session. The script checks `OneDriveCommercial` and `OneDrive` environment variables. Use `-Location Local` if OneDrive is unavailable. Verify internet connectivity. Some corporate networks may block downloads. Supported formats: ZIP and self-extracting EXE (7z syntax). The script configures registry compatibility settings automatically. If issues persist, verify the registry entry at: `HKCU:\Software\Microsoft\Windows NT\CurrentVersion\AppCompatFlags\Layers` You should see a registry Name similar to: `C:\Users\USERNAME\.local\bin\claude.exe` With a Data value of `~ ARM64HIDEAVX` --- title: "Claude Cowork \u2014 Overview and Security Considerations" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-cowork-overview-and-security-considerations.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-cowork-overview-and-security-considerations.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/836698117/Claude+Cowork+\u2014+Overview+and+Security+Considerations" page_id: "836698117" tags: ["claude", "overview", "security", "policy", "enterprise"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-02-20" --- --- ### What is Claude Cowork? Claude Cowork is a feature built into the Claude Desktop application that gives Claude the ability to act as an autonomous desktop agent. Rather than responding to prompts one message at a time like standard Claude Chat, Cowork can take on complex, multi-step tasks and execute them on your behalf reading, editing, creating, and organizing files directly on your computer. You point Cowork at a folder on your machine, describe the outcome you want, and Claude makes a plan and works through it step by step. Example use cases include assembling an expense report from a folder of receipt photos, reorganizing a messy downloads folder, drafting documents from source materials, or converting files between formats. Cowork is built on the same agentic architecture that powers Claude Code, but packaged in a more approachable interface for non-coding tasks. It launched in January 2026 as a research preview and is available on both macOS and Windows through the Claude Desktop app. --- ### How Does Cowork Differ from Claude Chat? In standard Claude Chat (what most Syracuse University users access at [claude.ai](http://claude.ai)), Claude responds to your messages conversationally. You upload files, ask questions, and get answers but Claude does not directly access or modify anything on your computer. Cowork is fundamentally different. When you grant Cowork access to a folder, Claude can autonomously read, write, create, and permanently delete files within that folder. It can also execute code, use browser extensions (Claude in Chrome), and connect to external services through plugins and MCP connectors. It operates more like a digital coworker than a chatbot. --- ### Current Status at Syracuse University **Cowork is currently disabled for all users in the Syracuse University Claude Enterprise organization.** We made this decision deliberately after evaluating the feature's security posture and its fit within our enterprise compliance requirements. While Cowork offers compelling productivity potential, several significant concerns need to be addressed before we can responsibly enable it for the campus community. --- ### Why Cowork Is Not Yet Enabled #### 1. No Audit Log or Compliance API Coverage This is the most critical concern. Anthropic has explicitly stated that Cowork activity is **not captured in Audit Logs, the Compliance API, or Data Exports**. Conversation history is stored locally on the user's computer only. For Syracuse University, this means we would have no centralized visibility into what Cowork is being used for, what files it accesses, or what actions it takes. This creates a significant gap in our ability to meet institutional compliance requirements, respond to incident investigations, or monitor for misuse. Anthropic's own documentation states: If your organization requires audit trails for compliance purposes, do not enable Cowork for regulated workloads. #### 2. Data Stored Locally, Not Subject to Enterprise Retention Cowork stores all conversation history locally on the user's machine. This data is not subject to Anthropic's standard data retention timeframe or our enterprise data governance policies. This means Cowork session data could persist indefinitely on a user's laptop (or be lost entirely if the machine is reimaged), with no organizational control over retention or deletion. #### 3. Local File System Access Creates Data Exposure Risk Cowork can read, write, and permanently delete files in any folder a user grants it access to. If a user inadvertently points Cowork at a folder containing sensitive university data student records, financial documents, credentials, research data that data could be processed by Cowork with no audit trail and no organizational visibility. The risk is amplified because users may not fully understand the scope of access they are granting. #### 4. Prompt Injection and Agentic Risk Because Cowork operates autonomously (reading files, browsing the web, executing code), it is susceptible to prompt injection attacks where malicious content hidden in a file or website attempts to alter Claude's behavior. Anthropic acknowledges this risk directly, noting that agent safety is still an active area of development. There have been documented real-world incidents, including cases where Cowork accidentally deleted files a user did not intend to be removed. #### 5. No Granular Admin Controls During the research preview, Cowork is controlled by a single organization-wide toggle on or off for everyone. There is no ability to enable Cowork for specific users, roles, or departments. Similarly, plugins installed within Cowork are saved locally to each user's machine and cannot be centrally provisioned or managed by admins. This all-or-nothing approach does not align with our need for role-based access and centralized policy management. #### 6. Research Preview Status Anthropic classifies Cowork as a research preview meaning the feature is still in active development, safety mechanisms are still being hardened, and the product may change significantly. Enabling a research preview feature across an enterprise organization of Syracuse's scale carries inherent risk, particularly given the other concerns listed above. --- ### What Can I Use Instead? Syracuse University users have access to a robust set of Claude capabilities that do not carry the same security concerns: - **Claude Chat** ([claude.ai](http://claude.ai)) Full conversational AI with file upload, artifacts, projects, web search, and connectors. Covered by enterprise audit logs and compliance controls. - **Claude Code** Terminal-based agentic coding tool for developers. Available via premium seat purchase. Covered by enterprise audit infrastructure. - **Claude API** Programmatic access for building applications and automations. Available via premium seat purchase. For more information on these products, see [Understanding Claude Products: Chat, Code, and API](./understanding-claude-products-chat-code-and-api.md). --- ### Questions? If you have questions about Cowork or other Claude features, please reach out: - **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) *Page last updated: February 2026* --- title: "Claude Enterprise at Syracuse University" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-enterprise-at-syracuse-university.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-enterprise-at-syracuse-university.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/534642749/Claude+Enterprise+at+Syracuse+University" page_id: "534642749" tags: ["claude", "overview", "getting-started", "enterprise", "productivity"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-30" --- ## **What is Claude Enterprise?** Claude Enterprise is Syracuse University's premium AI assistant that helps you work smarter, not harder. Think of Claude as your intelligent digital colleague who can help you write, analyze, research, and solve problems across virtually any subject area. [Sign-Up Now!](https://getclaude.syr.edu/) ### **Getting Started: The Basics** The easiest way to get started is to ask Claude a question or give Claude a task. Here are a few simple ideas: - **Writing Assistant**: Help draft emails, reports, essays, and creative content - **Research Partner**: Summarize articles, explain complex topics, and find connections between ideas - **Problem Solver**: Break down complicated problems, suggest solutions, and help with analysis - **Learning Companion**: Explain concepts in simple terms, create study guides, and answer questions - **Data Analyst**: Help interpret data, create summaries, and identify patterns - **Creative Collaborator**: Brainstorm ideas, develop proposals, and refine your work ### **How Does It Work?** Using Claude is as simple as having a conversation. You type your question or request in plain English (or other languages!), and Claude responds with helpful, detailed answers. No coding or technical knowledge required. **Example interactions:** - "Help me write a professional email declining a meeting invitation" - "Explain quantum computing like I'm a freshman" - "What are the key points in this 50-page PDF I just uploaded?" - "Help me prepare for my presentation on climate change" - "Create a study schedule for my finals week" ### **Best Practices for New Users** **Start Simple**: Begin with straightforward questions to get comfortable with how Claude responds. **Be Specific**: The more detail you provide, the better Claude can help. Instead of "help with my essay," try "help me create an outline for a 5-page essay about the impact of social media on democracy." **Iterate and Refine**: Don't hesitate to ask follow-up questions or request changes. You can say things like: - "Make this more formal" - "Explain that in simpler terms" - "Give me three more examples" - "Make this shorter" **Verify Important Information**: While Claude is highly capable, always double-check critical facts, citations, and data for important work. --- ## Showcase of Syracuse University Claude uses [Transform your meeting transcripts into actionable summaries, meeting minutes, and task lists using Claude](./meeting-summaries-with-claude.md). [Create a Study Project inside Claude to master subjects](./claude-study-project.md) [Create and revise resumes, draft personalized cover letters, practice behavioral interview questions](./claude-career-project.md) --- #### Additional Info [Claude Frequently Asked Questions](./claude-frequently-asked-questions.md) [Purchase Claude Code access](./purchase-claude-code-and-claude-api-access.md) --- title: "Claude - Frequently Asked Questions" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-frequently-asked-questions.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-frequently-asked-questions.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488210484/Claude+-+Frequently+Asked+Questions" page_id: "488210484" tags: ["claude", "faq", "data-privacy", "policy", "security"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-03-13" --- > [!info] > You may see Anthropic and Claude used interchangeably. Anthropic is the name of the company that owns and develops Claude. ## Do I retain ownership of my data I upload to the Claude platform? The Anthropic Claude terms of service that govern the Syracuse University instance of Claude state that the user retains all ownership of any inputs to the Claude platform and is granted ownership over any outputs that Claude creates. --- ## What classifications of data may be uploaded to the Claude platform when I logged in with my Syracuse University credentials to [approved AI tools](./approved-tools-for-use-with-university-data.md)? All [classifications](https://answers.atlassian.syr.edu/wiki/search?text=Data+Classification+Definitions&spaceKey=infosec011) of university data can be uploaded and used when a user is logged in with university credentials (NetID). Please be careful of the data you upload to a Claude Project or Claude Artifact if it is shared or plan to share it. --- ## How long does the Claude platform retain items like chat history, projects, artifacts, & items uploaded? Currently Claude retains all data for 2 years. When you delete a chat or project, it is no longer visible to you and will no longer be part of your personalized memories in the platform. All deleted items are retained in the platform for the duration of the retention period independent of visibility. --- ## What level of access does SU have to my AI usage in terms of data, prompts, and information input into the AI platforms? As with all computing systems, data stored within Syracuse University information systems is inherently accessible to authorized IT staff in order to support, secure, maintain, upgrade, troubleshoot, and back-up those systems. At Syracuse, this access is strictly governed by the Information Technology Resources Acceptable Use policy () and the Information Security Framework and is exercised in accordance with departmental guidelines and best practices. --- ## Does the Claude platform train on my data? By default the data that Syracuse University users upload and the chat sessions they create are not used to train any of the Claude platform AI models. If you explicitly report feedback or bugs to Claude, then Claude may use your submission to train AI models. The Claude platform has personalization features that are configurable by each user that allow for the platform to remember a user and use prior chat sessions and projects when generating new outputs. To change any memory or other personalization settings please visit the [Claude Settings](https://claude.ai/settings/features) page. More information can be found at: [Understanding Claude's Personalization Features | Anthropic Help Center](https://support.claude.com/en/articles/10185728-understanding-claude-s-personalization-features) --- ## Are Incognito chats saved? Incognito chats are temporary conversations that aren't saved to your chat history or to your personalized [Claude memory](https://support.claude.com/en/articles/11817273-using-claude-s-chat-search-and-memory-to-build-on-previous-context). The chat session is still retained on the Claude platform and follows the retention policy of all other chats and projects. --- ## How can I get access to Claude Code? Claude Code is a way to consume Claude models via API access and the Claude Code terminal. The Claude Enterprise License does not include Claude Code, but users can purchase access via a credit card. [Learn More](./purchase-claude-code-and-claude-api-access.md). --- ## Can I access Claude while traveling to other countries? Anthropic's restricts access from certain countries. Attempting to access Claude from a prohibited country may result in your Syracuse University Claude account being banned by Anthropic. If your account is banned, please see: For a complete list of countries where Claude is available, visit: [Anthropic | Supported Countries](https://www.anthropic.com/supported-countries) --- ## My SU Claude account was banned. Can you help? If your account was banned by Anthropic, you'll need to submit an appeal directly to them through their support system: **Appeal process:** When submitting your appeal: - Be honest and accurate about what happened - Review Anthropic's public [Terms of Service](https://www.anthropic.com/legal/consumer-terms) and [Usage Policy](https://www.anthropic.com/legal/aup) to understand what may have triggered the ban - Anthropic only allows access from some countries, see: [Anthropic | Supported Countries](https://www.anthropic.com/supported-countries) - Explain your intended use case clearly Note that appeals are reviewed by Anthropic's Trust Safety team, and approval is not guaranteed. While Syracuse University has an Enterprise agreement with Anthropic, we cannot influence their account ban decisions or appeal outcomes. --- ## What is the difference between Claude Haiku, Sonnet, and Opus? These are three different versions of Claude, each optimized for different needs: **Claude Haiku** - Fastest and most efficient. Best for quick, straightforward tasks. **Claude Sonnet** - Balanced performance and speed. Recommended for everyday use and most general tasks. **Claude Opus** - Most capable and intelligent. Best for complex tasks requiring advanced reasoning and analysis. Think of them as small, medium, and large options - choose based on the complexity of your task. **Note:** Larger models (Opus and Sonnet) consume more of your usage limits than Haiku. Consider using Haiku for simpler tasks to preserve your capacity for more complex work using Sonnet or Opus. --- ## How do I reset or extend Claude limits or access Claude Code? In general, Claude limits reset through the day. Requests submitted through the Claude interface asking for usage limit increases has been disabled and requests will **not** be granted outside of the processes outlined below. Claude Code and Claude API require a premium seat, more information found here: Anthropic has stated that they can change what limits are dynamically to best provide a stable environment. It is recommended you follow the [best practices](https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488210484/Claude+-+Frequently+Asked+Questions#What-are-some-best-practices-when-using-Claude%3F) when using Claude or other AI systems to not exhaust limits unnecessarily. Some additional information from Claude: **Staff and Faculty:** A Premium Seat does **not** remove all usage limits. It simply provides a higher limit. It is still possible to run out of usage even with Premium seat. A Premium seat provides: 1. Increased usage limits per month compared to standard access 2. Access to additional features including Claude Code and the Claude API Regarding usage limits: While Premium provides higher limits, it does still have usage caps. If you anticipate high usage that may exceed these limits, we can configure an overage billing option. This works as follows: 1. Once you reach your Premium usage limit, you can continue using Claude on a pay-as-you-go basis 2. Overage charges are calculated per-usage and billed to the payment method associated with your Premium seat 3. We can establish a spending cap to prevent unexpected large bills If you expect to require this overage billing configuration, please let us know a spending cap you are comfortable with and we can set it up for your account. Otherwise, your Premium package request should meet your needs. We would recommend you try using the Premium seat as-is first, and if you find yourself hitting the limits often and are open to spending more for usage, please let us know by emailing aihelp@syr.edu. > [!info] > You must have a Premium seat to have a usage based pay-as-you-go added on top. Syracuse Faculty and Staff can use university funds to purchase the Premium license. These funds will most likely originate from your department and may require approval from a department head. To start the process, please request the the [Claude Premium Access Package](https://myaccess.microsoft.com/@sumailsyr.onmicrosoft.com#/access-packages/9872d51f-fd45-4c44-8d6a-e2fd2975b818). **Students:** We do not currently provide premium seats for students. If you are a student worker and/or a department has authorized to pay for a premium seat for you, please have your supervisor and/or the department head reach out to us to set it up. This is subject to change, but the premium seat would be charged to the department at $2400/year, pro-rated from September 16th **I do not have approval for funding or do not wish to pay extra:** If you anticipate needing a higher usage limit and do not wish to pay for a Claude premium seat or credits, we would recommend you explore the platform to see if it can meet your needs. The mentorAI platform includes the Anthropic/Claude models. Usage limits in the mentorAI platform are higher and may offer tools Claude does not offer (screen sharing, etc) --- ## What are some best practices when using Claude? Claude offers several different models. Different models consume allocation at different rates. For example, Opus tier model uses **much more** of your allocation per prompt than the Sonnet model. Identify which model you *need* for your task. If the task is routine (e.g., summarizing text, minor edits), pick the lighter model. If its complex (e.g., large architectural reasoning, research summary), use the heavier model - but be aware a heavier model will use a larger part of your usage limit. [More information about Claude models in the above F.A.Q.](https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/488210484/Claude+-+Frequently+Asked+Questions#What-is-the-difference-between-Claude-Haiku%2C-Sonnet%2C-and-Opus%3F) **For Coding Tasks** - Provide complete context about your coding environment in your initial message. - Include entire relevant code snippets in one message for reviews or debugging. **For Writing Assistance** - Outline requirements, target audience, and key points comprehensively. - Send entire texts for editing in one message rather than breaking them up. **For Research and Analysis** - Clearly define your research question and focus areas initially. - Provide all relevant data in a single, well-structured message. By following these best practices, you can make the most efficient use of your Claude plan's message allocation. --- ## **Why can't Claude access the internet when running code?** When Claude executes code through its computer tools (bash, Python, R, etc.), network access is disabled. This is an intentional security configuration for Syracuse University's Claude Enterprise deployment. **Why is this restriction in place?** Claude's code execution environment has access to any data you bring into your conversationwhether through file uploads or connected services like Microsoft 365. Enabling network access would create a risk where code could potentially send sensitive University data to external endpoints. By keeping network access disabled, we ensure that any code Claude runs remains sandboxed and cannot transmit data outside the secure environment. **What can I do instead?** - **Download data first:** If you need to analyze data from an external source, download it to your computer and then upload it to Claude. - **Use Claude to write code:** Ask Claude to help you write scripts that you can then run in your own local environment (RStudio, Jupyter, VS Code, etc.) where you have full control over network access. - **Web search is still available:** Claude can still search the web and retrieve information conversationallythis restriction only applies to code execution. **Will this change in the future?** We continue to evaluate our security posture as Anthropic releases new features and controls. If a secure method for allowing limited network egress becomes available, we will consider enabling it. --- title: "Claude Local MCP - Connecting Claude Desktop to Power BI" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-local-mcp-connecting-claude-desktop-to-power-bi.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/claude-local-mcp-connecting-claude-desktop-to-power-bi.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/837517313/Claude+Local+MCP+-+Connecting+Claude+Desktop+to+Power+BI" page_id: "837517313" tags: ["claude", "how-to", "mcp", "integration", "setup"] audience: ["faculty", "staff", "IT"] last_modified: "2026-05-15" --- --- # **What is Local MCP?** Claude local MCP (Model Context Protocol) is an open-source standard developed by Anthropic that allows the Claude Desktop app to securely connect to, read, and interact with data and tools directly on your local computer, such as files, folders, databases, and APIs. It enables AI-driven automation, allowing Claude to perform actions like editing code, querying local data, or interacting with software tools without leaving your desktop interface. --- # Using Power BI MCP as an example 1. Download MCP from VS Code Search for the Power BI Modeling MCP extension from VS Code, and download the one published by Microsoft. ![image-20260220-192341.png](./attachments/837517313/image-20260220-192341.png) 1. Find the powerbi-modeling-mcp.exe file from the VS Code extension file. You can find it from a similar file path: ``` C:\Users\\.vscode\extensions\analysis-services.powerbi-modeling-mcp-0.1.9-win32-x64\server ``` ![image-20260220-192904.png](./attachments/837517313/image-20260220-192904.png) 1. Hold shift, right-click the exe file, and copy as path, save it somewhere for later use 2. Open the Claude desktop APP, click the edit config button as shown below ![image-20260220-193155.png](./attachments/837517313/image-20260220-193155.png) 1. Open the config file in VS Code, paste the configuration below, and save it ``` { "mcpServers": { "powerbi-modeling-mcp": { "command": "paste the path you copied before here", "args": ["--start"], "env": {} } } } ``` > [!info] > Remember to use double backslashes (\\) instead of single ones (\) in the path. 1. Reboot the computer, then open the Claude desktop APP. You should see the MCP added to the Claude. ![image-20260220-194005.png](./attachments/837517313/image-20260220-194005.png) > [!info] > For adding more mcps, you can use similar measure by using , separating them see example below ``` { "mcpServers": { "powerbi-modeling-mcp": { "command": "your path", "args": ["--start"], "env": {} } }, "mcpServers": { "other mcp": { "command": "your path", "args": ["--start"], "env": {} } } } ``` --- # How to use it? Now, you can open any Power BI report you are working on and open a chat from Claude Desktop. Input the prompt > *Connect to the open Power BI desktop file* From here, Claude will do its magic. ![image-20260220-195700.png](./attachments/837517313/image-20260220-195700.png) ![image-20260220-195824.png](./attachments/837517313/image-20260220-195824.png) > [!info] > You are in control of how much access Claude can have. --- title: "Connect Claude to M365" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/connect-claude-to-m365.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/connect-claude-to-m365.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/544210961/Connect+Claude+to+M365" page_id: "544210961" tags: ["claude", "how-to", "integration", "m365", "setup"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-02-24" --- ## Overview Claude can be connected to Microsoft 365 via the connector. This will allow your Claude instance to reference materials and gain context based on resources you may have in Microsoft 365. The Claude Microsoft 365 connector will allow Claude to access Microsoft Teams messages, Outlook emails, Sharepoint and OneDrive documents, and more. (if you grant this access when configuring) --- ## Video If you wish to watch a video of the process, please see the link below --- ## Process Follow the steps below to connect Microsoft 365 to Claude 1. Log into Claude under your Syracuse University account (*netid*@syr.edu) 2. Click your profile in the bottom left and select Settings ![A screenshot of the menu that appears when clicking the profile in the bottom left and a red box highlight the Settings option](./attachments/544210961/image-20251030-161518.png) 3. From here, press Connectors and next to the Microsoft 365 option press Connect ![Screenshot showing the Connectors option selected and the Microsoft 365 connector option](./attachments/544210961/image-20251030-161659.png) 4. Sign in, or select, your Syracuse University Microsoft account (*netid*@syr.edu) 5. You should be redirected back to Claude and have Microsoft 365 connected. 6. Feel free to press Configure next to the connector and change any options you desire or leave it configured as-is. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "AI Research Assistant with Claude" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/ai-research-assistant-with-claude.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/ai-research-assistant-with-claude.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/572194844/AI+Research+Assistant+with+Claude" page_id: "572194844" tags: ["claude", "example-use", "research", "teaching-learning"] audience: ["students", "faculty", "staff"] last_modified: "2025-11-10" --- Claude can act as your personal research assistant. Analyze sources, draft outlines, compare research, and explore ideasall in one place with enhanced citation support. ## Research Smarter with Claude 1. **Upload documents** (PDFs, Word docs, research articles) or paste text directly into Claudeup to 5 files per conversation. 2. **Start with a clear prompt** describing what you need (e.g., Can you summarize findings, extract key quotes, or explain theories in plain language?) 3. **Use Projects** to organize research by course or topicupload all your readings once and reference them across multiple conversations. 4. **Create Artifacts** for summaries, outlines, and literature reviews that you can edit, refine, and export. > [!info] > βœ… **Recommended:**Use Claudes *Projects* or *Chat with Files* features to store multiple readings, track notes, and refine ideas collaboratively. --- πŸ’¬ *Example prompts:* > [!note]- Prompt Article Summary > Summarize this research article in 34 bullet points. > Include: > > - The main research question or hypothesis > - The methods used > - Key findings and implications > - Limitations or future directions. > > Keep the tone objective and academic. > [!note]- Prompt Compare Two Sources > Compare these two articles on [insert topic, e.g., climate adaptation or social media behavior]. > > Identify similarities and differences in: > > - Research questions > - Methodologies > - Key conclusions > > Create a comparison table showing these differences side-by-side. > [!note]- Prompt Literature Review Outline > Create an outline for a short literature review on [insert topic] using the uploaded sources. > > Organize it by theme, list 34 main points under each theme, and note where research gaps exist. Format this as an editable document. > [!note]- Prompt Extract and Organize Quotes > Extract 5-7 key quotes from this article that support the main argument. > > For each quote, provide: > > - The quote itself > - Page number (if available) > - A brief explanation of its significance > - Which section of my paper it might support ![](./attachments/572194844/Screenshot 2025-11-10 at 2.27.20β€―PM.png) > [!info] > ### **Pro Tips** > > - Begin with Summarize this article, then follow up with Now focus specifically on the limitations section. > - Request Include specific page numbers or section references to make it easier to verify and cite properly. > - Always double-check quotes, data, and interpretations through SU Libraries or your instructor's recommended materials. Use Claude as a research aid, not a replacement for critical reading. > - Request specific formats like Create this as an APA-style annotated bibliography or Format this as a table. --- ## Claude's Unique Features for Research > [!note]- πŸ“ Projects (School Accounts) > Upload all your course readings, lecture notes, and research materials to a Project. Claude will remember and reference these materials across all conversations in that Projectno need to re-upload. > > **Example use:** Create a PSY 101 Project with your textbook chapters and articles, then ask questions throughout the semester. > [!note]- πŸ“‘ Multi-Document Analysis > Upload up to 5 documents at once and ask Claude to synthesize information across all of them. Perfect for comparing multiple sources or building comprehensive literature reviews. > > **Try:** I've uploaded three articles on neural networks. Create a synthesis matrix comparing their approaches, datasets, and conclusions. > [!note]- πŸ” Web Search Integration > Claude can search the web for current information when needed, providing citations to recent studies, news, or data. > > **Try:** Search for recent 2025 research on climate change adaptation strategies and summarize the key findings. > [!note]- 🧠 Extended Thinking > For complex research questions, Claude can use extended thinking to reason through problems more deeply, showing its analytical process as it works. > > **Try:** Use extended thinking to analyze the methodological strengths and weaknesses across these three studies. > [!note]- ✏️ Interactive Artifacts > Claude generates editable documents, tables, and outlines in a side panel. You can ask Claude to revise specific sections without starting over. > > **Try:** After generating a literature review outline, say Expand the methodology section with more detail from Article 2. > [!note]- 🎨 Use Styles (Custom Instructions) > Create custom styles that tell Claude how you prefer information formattedperfect for maintaining consistency across all your research notes and summaries. > > **Try:** Create a style like Always format summaries with IMRAD structure (Introduction, Methods, Results, and Discussion) and include methodology details. ### *Keep in Mind* - **Web search draws from public sources,** not SU's library databases. Always verify through SU Libraries for peer-reviewed sources. - **Claude is a research assistant, not a replacement for critical thinking.** Use it to organize, summarize, and explorebut form your own analysis. - **Always cite original sources.** Claude can help you find and organize information, but proper attribution is your responsibility. - **Extended thinking takes longer but goes deeper:** When you need thorough analysis of complex material, extended thinking provides more rigorous reasoningthough it requires extra processing time. - **Artifacts are collaborative:** Think of artifacts as living documents you can refine through conversation rather than static outputs. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Claude Career Project" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/claude-career-project.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/claude-career-project.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/511246346/Claude+Career+Project" page_id: "511246346" tags: ["claude", "example-use", "productivity"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-22" --- You can create and revise resumes, draft personalized cover letters, practice behavioral interview questions and research companies to prepare for internships or job applications. Brief overview: 1. Select Career chat or create a Career project. 2. Upload your resume and job description. 3. Ask Claude: *Tailor my resume for this role* or *Write a cover letter based on my profile.* Example: ![image-20251015-131744.png](./attachments/511246346/image-20251015-131744.png) ![image-20251015-131856.png](./attachments/511246346/image-20251015-131856.png) ![image-20251015-131934.png](./attachments/511246346/image-20251015-131934.png) ![image-20251015-131958.png](./attachments/511246346/image-20251015-131958.png) --- title: "Claude Study Project" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/claude-study-project.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/claude-study-project.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/511279124/Claude+Study+Project" page_id: "511279124" tags: ["claude", "example-use", "study-tools", "teaching-learning"] audience: ["students", "faculty"] last_modified: "2025-10-22" --- Create a *Study Project* inside Claude to master subjects. Upload lecture slides, PDFs, or notes and ask Claude to explain, quiz , summarize chapters or generate flashcards. How to use it: 1. Click on Learn or Study project from the home screen. 2. Upload class materials via the project setup. 3. Ask Claude: *Summarize this PDF into flashcards* or *Explain the key concepts from this slide.* Example: ![image-20251015-132147.png](./attachments/511279124/image-20251015-132147.png) ![image-20251015-132251.png](./attachments/511279124/image-20251015-132251.png) ![image-20251015-132658.png](./attachments/511279124/image-20251015-132658.png) --- title: "Drafting Emails with Claude" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/drafting-emails-with-claude.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/drafting-emails-with-claude.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/516325410/Drafting+Emails+with+Claude" page_id: "516325410" tags: ["claude", "example-use", "productivity"] audience: ["students", "faculty", "staff"] last_modified: "2025-11-11" --- Write clear, professional, and confident emails fast. Claude helps you compose and refine messages using natural prompts and built-in editing tools. ## Compose Smarter with Claude ### Steps 1. **Ask Claude to draft an email in an artifact** (e.g., Write a professional thank-you email to my professor about career advice. Put it in an artifact.) - **Important**: Add put it in an artifact or create this as an artifact to your prompt to ensure the email opens in a separate panel. Otherwise, shorter emails may appear inline in the chat. 2. **Your email appears in the Artifact panel** the draft displays in a separate, editable panel on the right side of the screen with easy copy/paste functionality. 3. **Refine your draft** with follow-up prompts like: - Make this more formal**.** - Shorten to 120 words. - Add a line about scheduling a meetin.g - Make the tone friendlier. 4. **Copy and send** Click the copy icon in the Artifact panel, then paste into your email client. --- πŸ’¬ *Example prompts:* > [!note]- Thank-You Email (Career or Networking) > Draft a professional thank-you email to [insert recruiter, alumnus, or interviewer's name] after [insert event, e.g., an informational interview or career fair conversation]. > > Include: > > - A sentence expressing gratitude for their time or advice > - One specific insight or topic discussed > - A short closing line about staying in touch or applying what you learned > > Keep the tone warm, professional, and under 120 words. Put it in an artifact. > [!note]- Email to a Professor > Write a polite and professional email to [insert professor's name] about [insert reason, e.g., requesting an extension on an assignment or clarification on course material]. > > Include: > > - A clear subject line (e.g., 'Follow-Up on MGT 248 Assignment') > - A respectful greeting > - A concise explanation of the situation > - A closing that expresses appreciation (e.g., 'Thank you for your understanding.') > > Keep the message under 150 words and use an academic, courteous tone. Put it in an artifact. > [!note]- Group Project Coordination > Compose an email to [insert group member names or 'my project team'] summarizing our [insert project name or presentation topic] next steps. > > Include: > > - A summary of what was completed > - Assigned tasks or roles for each member > - Our next meeting time or deadline > - A short, positive closing > > Keep the tone collaborative and friendly, around 100150 words. Put it in an artifact. > [!note]- Internship Follow-Up > Write a concise follow-up email to [insert recruiter's name or company] after submitting my application for the [insert internship title] position. > > Include: > > - A thank-you for reviewing my application > - 12 sentences connecting my coursework or skills to the role > - A polite closing showing continued interest > > Keep the tone professional, brief, and confident (around 120 words). Put it in an artifact. ![Screenshot of Claude AI interface showing a user requesting a professional thank you email to Dr Edward after a career fair. The left panel displays the prompt and Claude's response summary while the right panel shows the completed email draft in an artifact with subject line and formatted message content expressing gratitude and mentioning their shared interest in the Yankees.](./attachments/516325410/Screenshot 2025-11-07 at 3.46.19β€―PM.png) > [!info] > ### **Pro Tips** > > βœ… Be specific about audience and purpose (professor, recruiter, teammate) for better results. > > ✏️ Personalize the draft Claude provides strong structure, but add your own details and voice. > > πŸ’‘ Use iterative prompts After Claude drafts, say make it more concise or add enthusiasm to refine. > > πŸ”’ Always review before sending Claude doesn't know your specific schedule, attachments, or recipient details. > > πŸ“‹ Save frequently used emails Keep Claude's drafts in a document for templates you use often. ### Keep These in Mind - **πŸ“ Claude drafts, you decide**: Claude creates the structure and language, but you're responsible for reviewing accuracy, tone, and context before sending. - **🚫 Avoid sensitive data**: Never include confidential information, FERPA-protected student details, or personal identifiers when using Claude. - **πŸ’¬ Stay authentic**: Claude helps with structure and clarity but always personalize your message so it reflects your own voice. - **πŸ“§ No direct email integration**: Unlike some AI tools, Claude doesn't integrate directly with Gmail or Outlook. You'll copy and paste the final draft into your email platform. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Meeting Summaries with Claude" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/meeting-summaries-with-claude.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/example-uses/meeting-summaries-with-claude.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/500236296/Meeting+Summaries+with+Claude" page_id: "500236296" tags: ["claude", "example-use", "productivity", "how-to"] audience: ["students", "faculty", "staff"] last_modified: "2026-01-23" --- Transform your meeting transcripts into actionable summaries, meeting minutes, and task lists using Claude. ## Summarize with Claude Meeting Transcripts Project **Youll need:** A transcript file from Zoom (.txt) or Teams (.docx) > [!info] > **Pro Tips: Use the**[**Meeting Transcripts**](https://claude.ai/projects?search=Meeting+Transcripts&category=team)**Project** - The project has the prompts already in place and Claude is ready to assist. Use of the project is private and not shared with any other Claude users. **Steps:** 1. Get your transcript ([Zoom: Save Transcript](https://support.zoom.com/hc/en/article?id=zm_kb&sysparm_article=KB0063899); [Teams: Download from meeting details](https://support.microsoft.com/en-us/office/view-live-transcription-in-microsoft-teams-meetings-dc1a8f23-2e20-4684-885e-2152e06a4a8b)). 2. Open the Claude [**Meeting Transcripts Project**](https://claude.ai/project/0199e821-95e3-74d4-b2ec-826cf3711895) 3. Upload your transcript file to the chat 4. Tell Claude what format you need: - πŸ“‹ Formal Meeting Minutes - Professional summary with decisions and action items - πŸ“‘ Topic-Based Summary - Organized by discussion topics - βœ… Action Items Only - Just tasks, deadlines, and who's responsible - ⚑ Quick 5-Minute Summary - Brief key takeaways - πŸ’¬ Custom - Describe what you need in your own words 5. Get your summary - Claude will automatically format it for you --- Prefer to write your own prompts? See example prompts below: πŸ’¬ *Example prompts:* > [!note]- Prompt - Meeting Minutes (Formal) > Create concise meeting minutes from this transcript with the following sections: > > - Meeting Purpose: Brief overview of why the meeting was held > - Key Discussion Points: Main topics covered > - Decisions Made: Any conclusions or agreements reached > - Action Items: Tasks assigned with responsible parties (if mentioned) > - Next Steps: Follow-up plans or future meetings > > Keep the summary professional and under 500 words. > [!note]- Prompt - Topic-Based Summary > Organize this meeting transcript by topic. For each topic discussed: > > - Provide a brief summary (2-3 sentences) > - List any decisions or conclusions > - Note any follow-up tasks or deadlines mentioned > > Format it as a bulleted list for easy scanning. > [!note]- Prompt - Action Items Tasks Only > Extract all action items, tasks, and deliverables from this transcript. For each item include: > > - The task description > - Who is responsible (if mentioned) > - Any deadlines or timeframes mentioned > - Priority level (if indicated) > > List them in order of urgency or deadline. > [!note]- Prompt - Quick 5-Minute Summary > Provide a brief 5-minute read summary of this meeting including: > > - 3-5 key takeaways > - Most important decisions > - Critical action items > > Keep it concise and scannable. ![Claude interface showing a .txt transcript file uploaded and the prompt 'Create concise meeting minutes with Purpose, Highlights, Decisions, and Action Items.'](./attachments/500236296/Screenshot 2025-10-08 at 11.19.23β€―AM.png) > [!info] > Pro Tips for Better AI Summaries > > - Be specific in your prompts: Tell the AI exactly what format or information you need > - Ask follow-up questions: Can you make this more concise? or Add deadlines if mentioned > - Edit the output: AI summaries are a starting pointreview and adjust for accuracy > - Save your favorite prompts: Create a prompt library for recurring meeting types --- ## FAQ Limitations > [!note]- Can Claude transcribe live meetings? > No. Claude can only summarize existing transcript text it cannot transcribe live audio or video. > > You must first generate a transcript using: > > - Zoom's built-in transcription > - Microsoft Teams transcription > - Or another transcription tool > > Then upload that text transcript to the AI tool for summarization. > [!note]- Can I upload audio or video files directly to AI tools? > Some AI tools can process audio files, but text transcripts are simpler and more reliable. > > Best practice: Use the text transcripts generated by Zoom or Teams, which already have consent notifications built-in. > [!note]- Does Microsoft Copilot for Work do transcription summaries? > Yes, Microsoft Copilot for Work can summarize meeting transcripts and even generate summaries automatically during Teams meetings. > > However, Copilot for Work requires an additional license and has to purchased through your IT Support. > [!note]- Are Zoom breakout rooms transcribed? > No. Zoom does not transcribe breakout room conversations with Syracuse University accounts. Only discussions in the main meeting room are included in automatic transcripts. > > What you can do instead: > > - Assign a note-taker in each breakout room to manually capture key points > - Have groups report back to the main room - these verbal summaries will be captured in the transcript > - Stay in the main room for discussions you need transcribed Keep These in Mind: - AI summaries may miss nuance or contextalways review the output - Transcription accuracy depends on transcript quality - AI tools have limits on file size and text length (though most meeting transcripts fit) - For sensitive content, use only approved AI tools per university policy When in doubt about data policies, contact ITS or review the AI Guidelines linked above. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Purchase Claude Code and Claude API Access" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/purchase-claude-code-and-claude-api-access.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/purchase-claude-code-and-claude-api-access.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/540934169/Purchase+Claude+Code+and+Claude+API+Access" page_id: "540934169" tags: ["claude-code", "access", "api", "setup", "enterprise"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-04-15" --- ## Overview Claude Code and Claude API are advanced features offered by Anthropic. Members of the Syracuse University community are provided standard seats/licenses for Claude Enterprise. These advanced features will incur a cost on the department and/or student if they wish to use them. > [!info] > Claude Code and Claude API are packaged together, so getting access to Claude Code will also provide access to Claude API. --- ## Faculty and Staff Information Faculty and Staff that wish to have access to Claude Code and/or Claude API will need to get approval from their department to purchase a premium seat license or use a university business credit card to purchase credits. ### Paying for Claude Code usage Claude Code can be purchased using Departmental Funds and is charged based on monthly usage to a university chart string. Fill out the request form at the link below to get setup with access. [Claude Code API Interest and Access Request Form](https://support.atlassian.syr.edu/servicedesk/customer/portal/791/group/1727/create/3601) You will be asked to provide the following information: | **Form Field** | **Description** | | --- | --- | | Monthly Spending Cap (USD) | The maximum amount your department has approved for monthly API credit charges, up to a maximum of $200 per month. If more than $200 is needed, please consider a premium seat license instead. | | Chart-string | The Fund 11 chart-string approved for monthly API credit charges. | | Financial Approver | Email address of the department financial approver. | | IT Lead | Email address of the department IT lead. | Once the request has been processed you will receive an email with an invitation to the Syracuse University Claude Code platform instance. Open the invitation email from Anthropic and click the acceptance link. Sign in with your Syracuse University credentials (`netid@syr.edu`). ### Purchasing Credits If you instead wish to use a university business credit card to just buy credits, please follow the process below and use the card to purchase the credits. > [!note] > Upon separation with the university, your Syracuse University Claude Enterprise access will be removed and with it, you will lose any unused credits. Be cognizant of your usage as to not over purchase credits. > [!note]- Process for Faculty and Staff - Business credit card > - Navigate to > - Choose Individual option on the create screen (when logged in with your SU account, you are still being protected by SUs contract even though you are purchasing the credits) > > ![Screenshot of the account selection screen for claude. Showing Individual and Organization options.](./attachments/540934169/image-20251029-195123.png) > > - Click on the Buy Credits button and complete the purchase of some amount of credits > > ![Screenshot of the home page once logged into Claude where there is a button to buy credits](./attachments/540934169/image-20251030-150512.png) > > - Finish the setup of Claude Code by following the [Set up Claude Code - Claude DocsSet up Claude Code - Claude Docs](https://docs.claude.com/en/docs/claude-code/setup) ### Premium Seat If you wish to have a Claude Premium seat and have approval from your department to fund it, navigate to the [Access Package](https://myaccess.microsoft.com/@sumailsyr.onmicrosoft.com#/access-packages/9872d51f-fd45-4c44-8d6a-e2fd2975b818) to request the premium seat. Following approval, the cost is pro-rated and billed to the department based on our contract with Anthropic. If you are encountering problems with rate limits, you may wish to have a premium seat purchased instead of buying credits, as it will increase those limits. --- ## Student Information ### Students Students will need to purchase Claude Code access by signing up for an Individual Claude Console Workspace and buying credits with their own funds, i.e. on a credit card. When purchasing access through your Syracuse University Claude Enterprise account, as opposed to your own account, you will have the benefits of being protected by Syracuse Universitys enterprise agreement with Anthropic. This agreement will help protect your data and not have it used for training, in addition to allowing you to upload university data into your Claude instance while following the Syracuse University guidelines. > [!note] > Upon graduating or leaving the university, your Syracuse University Claude Enterprise access will be removed and with it, you will lose any unused credits. Be cognizant of your usage as to not over purchase credits. ### Student Workers/Graduate Assistants If you are a student worker or graduate assistant who feels they need Claude Code for their work, speak with your department about the option to have them purchase a premium seat or get you credits. The department can then reach out to ITS for assistance in this process. ### Process > [!note]- Process for Student Access > - Navigate to > - Choose Individual option on the create screen > > ![Screenshot of the account selection screen for claude. Showing Individual and Organization options.](./attachments/540934169/image-20251029-195123.png) > > - Click on the Buy Credits button and complete the purchase of some amount of credits > > ![Screenshot of the home page once logged into Claude where there is a button to buy credits](./attachments/540934169/image-20251030-150512.png) > > - Finish the setup of Claude Code by following the [Set up Claude Code - Claude DocsSet up Claude Code - Claude Docs](https://docs.claude.com/en/docs/claude-code/setup) --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Requesting a Claude Connector" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/requesting-a-claude-connector.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/requesting-a-claude-connector.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/841875458/Requesting+a+Claude+Connector" page_id: "841875458" tags: ["claude", "integration", "mcp", "access", "how-to"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-02-23" --- Claude Enterprise supports connectors integrations that allow Claude to read from and interact with third-party tools and services you use at work. Some connectors (like Microsoft 365 and Atlassian) are already enabled for all SU users. Others require a review and approval before ITS can enable them university-wide. ## What Is a Claude Connector? Connectors extend Claude's capabilities by giving it access to external systems via the Model Context Protocol (MCP). For example, the Microsoft 365 connector lets Claude search your SharePoint files and Teams content. The Atlassian connector allows Claude to read Jira tickets and Confluence pages. Connectors are configured at the organization level by ITS administrators. Once enabled, individual users can connect their personal accounts and toggle connectors on or off within their Claude settings. ## How to Request a New Connector If a tool you use has a Claude connector and you would like ITS to enable it, submit a request to [aihelp@syr.edu](mailto:aihelp@syr.edu) with the following: 1. The connector name and a link to the vendor's connector or MCP documentation 2. Your business use case describe how you plan to use the integration and what problem it solves 3. Confirmation of an SU contract or agreement with the vendor, if applicable 4. A SOC 2 Type II report from the vendor, or confirmation that you can obtain one ## What We Review Before Approving a Connector ITS evaluates connector requests based on several key factors. **Security Compliance:** The most important requirement is that the third-party vendor can provide a SOC 2 Type II audit report. This confirms the vendor has adequate security controls for handling university data. Many vendors publish this on a trust page or you can request it from your account representative. Requests will be on hold until this documentation is provided. **Existing University Relationship:** Connectors for tools where SU already has a contract are easier to approve. If your department uses a tool through a university-wide agreement, that supports the case for enabling its connector. **Data Classification Risk:** We consider what kind of university data may flow through the connector. Connectors that could expose Confidential data such as FERPA records, HIPAA data, PII, or financial information face a higher bar for approval and may require additional security review. **Breadth of Use:** Because connectors are enabled at the organization level, ITS weighs whether a connector benefits a broader population of SU users versus a single individual or team. Highly specialized tools with limited campus use may be deprioritized. **Vendor Maturity:** We look at whether the vendor's MCP or connector implementation is well-supported. Early-stage or experimental integrations may be held until they stabilize. ## What Could Prevent Approval A request is unlikely to be approved if the vendor cannot provide a SOC 2 Type II report, if SU has no existing contract with the vendor, if the tool is not approved for use with university data, if the connector would expose restricted data without adequate controls, or if the functionality is already available through an existing approved connector. ## Frequently Asked Questions **Can I use a connector that is not on the approved list?** Not through the university Claude instance. Connectors must be reviewed and enabled by ITS before they appear in your Claude settings. Submit a request to [aihelp@syr.edu](mailto:aihelp@syr.edu). **How long does the review take?** It depends on how quickly vendor documentation can be obtained. Simple requests where a SOC 2 report is readily available may resolve within a few business days. More complex reviews take longer. **What if my vendor does not have a SOC 2 report?** ITS will work with you to assess alternatives, but in some cases the request may not be approvable under current policy. **Can I use connectors in Claude Desktop that are not enterprise-enabled?** Claude Desktop supports local MCP connections that run entirely on your machine without transmitting data to a third-party server. These local integrations may be appropriate for certain use cases. ## Questions? Reach out to the AI team at [aihelp@syr.edu](mailto:aihelp@syr.edu) or book a consultation at the AI at Syracuse University Bookings page. --- title: "Understanding Claude Products: Chat, Code, and API" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/understanding-claude-products-chat-code-and-api.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/understanding-claude-products-chat-code-and-api.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/522158118/Understanding+Claude+Products:+Chat,+Code,+and+API" page_id: "522158118" tags: ["claude", "claude-code", "overview", "api", "access"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2025-11-10" --- --- ### What is Claude Chat? Claude Chat is the conversational interface you use to talk with Claude directly. It's available through the web at , as well as mobile apps (iOS and Android) and desktop applications. This is where you can have natural conversations with Claude for everyday tasks like writing, analysis, brainstorming, coding help, research, and more. > [!note]- What is Claude Chat useful for? > - Quick questions and conversations with Claude > - Writing assistance (emails, documents, creative content) > - Creating and editing content with Artifacts > - Creating Projects for specialized workspaces > - Analysis and research > - Learning new topics > - Brainstorming and ideation > - Getting coding help and explanations > - General productivity tasks #### **How do I request access to Claude Chat?** Claude Chat is available to everyone at . You can create an account, following the instructions in . Mobile and desktop apps are also available for download. --- ### What is Claude Code? Claude Code is a command-line tool that brings Claude into your terminal for agentic coding workflows. Instead of copying code back and forth between a chat interface and your editor, Claude Code works directly within your local development environment and can autonomously handle coding tasks. > [!note]- What is Claude Code useful for? > - Complex refactoring across multiple files in your codebase > - Implementing new features with minimal direction > - Debugging and troubleshooting in your actual development environment > - Running tests and fixing issues > - Working autonomously on coding tasks within your project > - Making systematic changes across your repository #### **How do I request access to Claude Code?** To learn more about Claude Code, visit --- ### What is Claude API? The Claude API provides programmatic access to Claude through Anthropic's developer platform. This allows developers to integrate Claude's capabilities directly into their own applications, tools, and services. You can access Claude Sonnet 4.5 via the API using the model string `claude-sonnet-4-5-20250929`. > [!note]- What is Claude API useful for? > - Building custom applications powered by Claude > - Creating automated workflows and integrations > - Scaling AI capabilities across your organization > - Developing tools that need Claude's language understanding > - Embedding conversational AI into existing products > - Processing large volumes of text or data programmatically #### **How do I request access to the Claude API?** To get started with the Claude API: 1. First, you will need to purchase credits or request a premium seat 2. Visit to access the Claude Developer Platform 3. Create an account to generate your API key 4. Follow the quickstart guide to integrate Claude into your application --- ### **Where can I learn more?** - **API Documentation**: - **Developer Platform Guide**: - **Claude Console**: Test and craft prompts in your browser at - **Cookbooks**: Practical code examples at - **Quickstarts**: Sample applications at - **Support**: --- ### Support Need help? ITS Help Desk: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Working with SharePoint Files in Claude" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/working-with-sharepoint-files-in-claude.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/claude/working-with-sharepoint-files-in-claude.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/988774401/Working+with+SharePoint+Files+in+Claude" page_id: "988774401" tags: ["claude", "how-to", "integration", "m365", "setup"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-04-16" --- Claude can find, read, and edit files stored in SharePoint including adding new entries, updating content, and creating new files. The key is connecting your SharePoint folder to your computer via a**OneDrive shortcut**, then granting Claude access to that folder using the **Filesystem connector** in the Claude Desktop app. > [!info] > The Filesystem connector requires the **Claude Desktop app**. If you only use Claude in the browser, you can dowload the desktop app via . ## Filesystem Connector Capabilities | **Action** | **Filesystem Connector Capability** | | --- | --- | | Read files in your shortcutted SharePoint folders | βœ… | | Find files by name in shortcutted folders | βœ… | | Edit or update existing files | βœ… | | Create new files | βœ… | | Delete files | ❌ | | Search across all SharePoint sites (not just shortcutted ones) | ❌ | | Access Outlook email, calendar, and Teams | ❌ | | Works in the browser | ❌ | ## Step 1: Add Your SharePoint Folder as a Shortcut in OneDrive - Open your **SharePoint site** in a web browser and navigate to the folder that has the files you want Claude to work with. - Click the **Add shortcut to My files** button in the toolbar at the top of the page. - You'll see a confirmation message. You can find it in **File Explorer** (Windows) or **Finder** (Mac) under your **OneDrive - Syracuse University** folder. It will have a small shortcut arrow on the icon. ![Screenshot 2026-04-15 162826.jpg](./attachments/988774401/Screenshot 2026-04-15 162826.jpg) ## Step 2: Enable the Filesystem Connector in Claude Desktop - Open the **Claude Desktop app**. If you haven't installed it yet, download it at and sign in with your SU credentials. - Find **Filesystem** in the connector list, then install and enable it. ![image.png](./attachments/988774401/image.png) - When configuring, **paste the**SharePoint shortcut path in the directory path. This is usually something like `C:\Users\YourName\OneDrive - Syracuse University` on Windows or `~/OneDrive - Syracuse University` on Mac. You can select the whole OneDrive folder or just the specific shortcut folder. You can add multiple paths. ![image (1).png](./attachments/988774401/image (1).png) Give it a try, Claude can now read and edit SharePoint files for you. --- ## Questions? Reach out to the AI team at [aihelp@syr.edu](mailto:aihelp@syr.edu) or book a consultation at the AI at Syracuse University Bookings page. --- title: "Clementine Class Search" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/clementine-class-search.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/clementine-class-search.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/895451142/Clementine+Class+Search" page_id: "895451142" tags: ["clementine", "overview", "teaching-learning", "faq", "getting-started"] audience: ["students", "faculty", "staff"] last_modified: "2026-04-01" --- # Clementine: Class Search Clementine: Class Search is Syracuse University's AI-powered class search tool built to help students find the right classes faster, in natural language. Instead of navigating complex catalogs or filtering through dropdowns, you just ask. Clementine understands natural language and searches across classes, descriptions, requirements, and more to surface what you're actually looking for. ### [Watch Clementine in Action](https://video.syr.edu/media/t/1_ckaodmmb) ### [Try Now @ clementine.syr.edu](https://clementine.syr.edu/) > [!note] Embedded iframe β€” Clementine Video > --- ## What Can Clementine Do? - **Search classes by topic, interest, or need** Ask things like show me writing-intensive classes in the humanities or what upper-division IST classes are available this fall? - **Find classes that meet specific requirements** Filter by credits, career level (undergraduate/graduate), academic term, or subject area - **Describe a class in plain language** Get summaries and details without digging through the course catalog - **Help you plan ahead** Explore options before your advising appointment or registration opens --- ## Who Is Clementine For? Clementine is available to all **Syracuse University students** and is especially useful during: - **Pre-registration planning** Explore your options before seats fill up - **Advising prep** Come to appointments with a shortlist already in hand - **Registration season** Quickly find classes that fit your schedule and requirements Academic advisors can also use Clementine to assist students during appointments. --- ## How to Access Clementine Clementine: Class Search is available at: [**clementine.syr.edu**](https://clementine.syr.edu) Select the Class Search Assistant. Log in with your Syracuse University credentials if prompted. --- ## **Queuing** To ensure a stable and reliable experience during periods of high demand, Clementine utilizes a queue system. If a large number of Syracuse University community members are accessing Clementine simultaneously, they may be placed in a queue before their session begins. When this occurs, a message will be displayed indicating their current position in the queue. We appreciate your patience and thank you for your understanding. --- ## Tips for Getting the Best Results - **Be specific** The more context you give, the better the results. 3-credit fall electives in data science works better than just data science - **Try different phrasings** If your first search doesn't land, rephrase and try again - **Use it as a starting point** Clementine helps you explore; always verify final enrollment details in MySlice or with your Academic Advisor. --- ## Frequently Asked Questions **Is Clementine the same as the course catalog?** No. Clementine is a conversational search layer on top of course data. It's designed to help you *find* classes, not replace the official catalog or MySlice for enrollment. **Does Clementine show real-time seat availability?** Clementine surfaces course information from the university's academic data. For live seat counts and registration, use MySlice. **Can Clementine tell me if a course counts toward my major requirements?** Clementine can help you find classes that align with requirements by description or attribute, but for official degree audit and requirement verification, consult your academic advisor or use Degree Works. **What makes Clementine different from Claude or other AI solutions?** Clementine is a purpose-built tool that leverages multiple technologies, including Claude, to deliver accurate and detailed course information specific to Syracuse University. The underlying data is sourced directly from the university's own datasets, meaning the depth and specificity of information available through Clementine cannot be found through general AI solutions or other external tools. **Who built Clementine?** Clementine was developed by the ITS Data AI team and the Registrars Office at Syracuse University as part of the university's broader AI innovation initiative. **Is Clementine an authoritative source?** No. Although Clementine is supported by a comprehensive dataset of course information, it is still an AI-powered tool and is subject to error. Users should verify critical information through official university channels. **Do Academic Advisors or Faculty Advisors have access to Clementine?** Yes. Clementine is available to all members of the Syracuse University community, including Academic Advisors and Faculty Advisors. **Can Academic Advisors or Faculty Advisors view what I have asked Clementine?** At this time, Academic Advisors and Faculty Advisors do not have access to a student's Clementine conversation history. **How does ITS handle the information I enter into Clementine?** For detailed information on how your data is managed, please visit our AI Data Privacy page: --- ## Need Help? For questions or feedback about Clementine, contact the ITS Service Center: - πŸ“§ [help@syr.edu](mailto:help@syr.edu) - πŸ“ž 315-443-2677 - 🌐 [its.syr.edu](https://its.syr.edu) --- title: "mentorAI Creating A Mentor" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-creating-a-mentor.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-creating-a-mentor.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/535068673/mentorAI+Creating+A+Mentor" page_id: "535068673" tags: ["mentorai", "how-to", "teaching-learning", "study-tools"] audience: ["faculty", "staff"] last_modified: "2025-11-11" --- Creating a mentor profile on MentorAI is a straightforward process that involves entering essential information, selecting visibility options, and configuring interaction prompts. This allows mentors to effectively connect with students seeking guidance. Please Reference for More mentorAI Guidance: [mentorAI Settings & Options](./mentorai-settings-options.md) [mentorAI Setting Up a Web Crawler](https://answers.atlassian.syr.edu/wiki/search?text=mentorAI+Setting+Up+a+Web+Crawler&spaceKey=ITSAI) ## Instructions 1. Click on the New Mentor button in your dashboard ![mentorAI home screen, highlighting the 'New Mentor' option on the left side of the screen. Under 'Mentors' and above 'New Project.'](./attachments/535068673/Screenshot 2025-11-11 at 1.44.43β€―PM.png) 1. **Input Mentor Information** 1. **Name** - **Enter the Mentor's Name:** Fill in the name of the mentor you want to create. - **Purpose:** This will be the primary identifier for mentees looking to connect. 2. **Description** - **Write a Brief Bio:** Provide a short description of the mentors background, skills, and what they can offer to mentees. - **Focus:** Include key experiences and the type of mentorship they provide (e.g., career advice, skill development). 3. **Category** - **Select a Category:** Choose the appropriate category for the mentor based on their expertise (e.g., Astronomy, Art, Anthropology). - **Importance:** This helps mentees find mentors relevant to their interests. 4. **Mentor Visibility** - **Choose Visibility Settings:** Decide whether the mentor will be visible to all students, select groups, or kept private. - **Consideration:** Think about how accessible you want the mentor to be for potential mentees. 5. **Image** - **Upload a Profile Picture:** If applicable, upload a professional image of the mentor. - **Tip:** A picture can enhance engagement and make the profile more relatable. !['Create Mentor' Settings screen showing options for 'Name,' 'Description,' 'Category,' and 'Image.'](./attachments/535068673/Screenshot 2025-11-11 at 1.45.00β€―PM.png) 1. **Configure Prompts** 1. **System Prompts** - **Review Existing Prompts:** These are usually pre-populated with suggestions for how the mentor should interact with mentees. - **Editing:** Feel free to modify these prompts to better fit the mentors style or focus areas. 2. **Proactive Prompts** - **Add Engagement Prompts:** These prompts can be used to initiate discussions with mentees. - **Customization:** Tailor these prompts to encourage specific types of interactions. 3. **Guided Prompts** - **Set Guidance for Conversations:** These prompts guide the mentors responses in a structured manner. - **Tip:** Make them relevant to the mentors expertise and the mentees needs. !['Create Mentor' Prompts screen showing options for 'System Prompts,' 'Proactive Prompts,' and 'Guided Prompts.'](./attachments/535068673/Screenshot 2025-11-11 at 1.45.17β€―PM.png) 1. Click the Save button to submit the new mentor profile. > [!info] > Add your configured prompts at the bottom of the default for more effective mentoring. --- title: "mentorAI Settings & Options" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-settings-options.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-settings-options.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/567279621/mentorAI+Settings+&+Options" page_id: "567279621" tags: ["mentorai", "how-to", "study-tools", "data-privacy", "integration"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2026-03-30" --- Use **Settings & Options** to customize how your mentorAI looks, behaves, and interacts with users. Each setting helps tailor your mentors purpose, audience visibility, and engagement style. ## 🧭 **Edit Mentor Overview** The **Edit Mentor** panel is where you manage and customize every aspect of your mentorAI. Each tab focuses on a different part of your mentors behavior, visibility, or capabilities from basic setup to advanced integrations. **Available Tabs:** - **βš™οΈ Settings:** Define your mentors name, description, category, and visibility permissions. - **🧠 LLM:** Choose which AI model your mentor uses and fine-tune its intelligence level. - **πŸ’¬ Prompts:** Configure default, proactive, and guided prompts that shape how the mentor interacts with users. - **πŸ›‘οΈ Safety:** Adjust moderation filters and set behavior limits to ensure safe, respectful interactions. - **❗ Disclaimers:** Add custom messages or disclaimers for transparency and ethical guidance. - **πŸ”§ Tools:** Connect your mentor to additional functions or APIs to enhance its capabilities. - **🧩 Memory:** Manage what information your mentor retains or recalls during conversations. - **πŸ•“ History:** View and manage recent interactions for monitoring or improvement. - **πŸ“Š Datasets:** Upload or connect data sources your mentor can reference for more tailored responses. - **πŸ–₯️ Embed:** Generate code to embed your mentorAI directly into web pages or digital platforms. --- ## **βš™οΈ Settings** ![mentorai settings page showing example fields for name description category and permissions](./attachments/567279621/Screenshot 2025-11-10 at 1.04.00β€―PM.png) | **Setting Options** | **Functions, Examples Tips** | | --- | --- | | **🧠 Name** | **What it does:** Defines the mentors display name this is how users will identify and search for it. **Example:** *Research Innovation MentorAI* **Tip:** Choose a name that clearly communicates the mentors focus area or purpose. | | **πŸ—’οΈ Description** | **What it does:**Outlines what the mentor does, who it supports, and how it helps. **Example:**Supports students in developing creative, well-structured academic and project ideas **Tip:** Write in a short paragraph, focusing on the mentors *function* and *value* to students. | | **🧭 Category** | **What it does:**Groups your mentorAI into a specific domain or topic area, like AI Assistance, Advisor or Artificial Intelligence. **Example:** *AI Assistance* for AIs supporting research, tech, or data-related questions. | | **πŸ‘οΈ Who Can View?** | **What it does:**Controls which users can see this mentorAI in the directory. **Options include:** - **Students:** Visible only to current students - **Administrators:** Visible to staff and admin users - **Anyone:** Publicly accessible | | #### πŸ’¬ Who Can Chat? | **What it does:** Determines who can start conversations with your mentorAI. **Options:** - **Authenticated Users:** Requires campus login (safest option for university AIs) - **Anyone:** Open access (ideal for public-facing mentors) | | **πŸ’Ύ Save** | **What it does:**Saves all entered information. **Note:** Make sure each required field (marked with a red asterisk *) is filled in before saving. | --- ## **🧠 LLM** The **LLM tab** lets you choose the underlying AI model that powers your mentorAI. Different providers offer their own models such as **Google's** **Gemini**, **OpenAIs** **ChatGPT**, and **Anthropics Claude** each with unique strengths and response styles. > [!note]- βš–οΈ Why Choose Different LLM Types > | **πŸ€– ChatGPT (OpenAI)** | Great for structured reasoning, academic QA, and professional dialogue. Provides balanced performance and clear communication. | > | --- | --- | > | **🌐 Gemini (Google)** | Excels in research, data interpretation, and multimodal reasoning (text + image). Ideal for analytical or innovation-based mentors. | > | **πŸ“š Claude (Anthropic)** | Strong at long-form writing, nuanced reflection, and summarization. Useful for writing support, analysis, or ethics-focused mentors. | > [!note]- 🧩 Tips for Syracuse MentorAI Use > - Choose the LLM that fits the **style and purpose** of your mentor. > - You can switch between models at any time to explore different response styles and performance levels. > [!info] > The LLM determines how your mentorAI thinks and communicates. Different models may vary in tone, creativity, and depth of reasoning try different ones to see what works best for your use case. --- ## **πŸ’¬ Prompts** ![mentorai prompts panel showing system proactive and guided prompts for research and innovation mentor](./attachments/567279621/Screenshot 2025-11-10 at 1.04.08β€―PM.png) | **🧱 System Prompt** | **⚑ Proactive Prompt** | **🧭 Guided Prompt** | | --- | --- | --- | | **Use:** Establishes the mentors overall purpose and tone. Think of it like a job description for the mentor. | **Use:** A short, automatic message that welcomes or engages users when a chat begins. | **Use:** Offers ready-made suggestions for users to click and start a conversation. | | **Use this for:** Describing the mentors role and how it should support users. | **Use this for:** Inviting users to ask questions or share what they need help with. | **Use this for:** Common actions or student questions such as Help me plan my essay or Explain this concept. | | **Tip:** Keep it focused on what the mentor can do for the intended audience. | **Tip:** Keep it friendly, short, and open-ended. | **Tip:** Add prompts that match common campus topics or student tasks. | **πŸ’¬ Suggested Prompts**appear as *conversation starters* at the bottom of your mentorAI chat window. These help users begin a conversation faster by suggesting common or useful questions related to the mentors topic. > [!info] > **πŸ’‘ Tip:**Keep suggested prompts short and action-oriented they work best when they feel like questions the mentors intended audience would naturally ask (e.g., Find sources for my topic or Help me summarize this reading). --- ## **πŸ›‘οΈ Safety** The **Safety tab** allows you to set moderation and safety rules for your mentorAI. These settings make sure that all conversations stay appropriate, inclusive, and aligned with Syracuse Universitys standards for respectful communication. ![mentorai safety panel showing moderation and safety prompts with example responses for research and innovation mentor](./attachments/567279621/Screenshot 2025-11-10 at 1.04.19β€―PM.png) #### βš–οΈ Moderation Safety Settings Each safety feature works together to filter, flag, and respond to inappropriate or sensitive interactions: > [!note]- Moderation Prompt Reponse > - **Moderation Prompt:** > Describes what types of *user messages* should be flagged such as hate speech, discrimination, or personal attacks. > - **Moderation Response:** > The message shown to users if their input violates guidelines. It reminds them to stay respectful and may include campus contact info (e.g., *Barnes Center at 315-443-8000* for confidential support). > [!note]- Safety Prompt Reponse > - **Safety Prompt:** > Outlines what types of *AI responses* should be restricted for example, messages that provide medical, legal, or financial advice. > - **Safety Response:** > The message displayed when the AI blocks an unsafe or off-topic question, encouraging users to seek help from appropriate campus resources instead. > [!info] > **πŸ’‘ Tip:**Keep both prompts and responses professional, concise, and student-focused. Together, they create a balanced safety system filtering inappropriate messages while redirecting students to the right campus support when needed. --- ## ❗ Disclaimers The **Disclaimers tab** allows you to add advisory messages that appear above the chat window. These short notices remind users to review responses critically and understand the AIs limitations. ![mentorai disclaimer panel showing advisory message about reviewing ai responses](./attachments/567279621/Screenshot 2025-11-10 at 1.04.26β€―PM.png) --- ## **πŸ”§ Tools** The **Tools tab** lets you manage what additional resources your mentorAI can access while responding to users. Find additional information here: [mentorAI - Tools](./mentorai-tools.md) ![mentorai tools panel showing trained documents and web search options turned on](./attachments/567279621/Screenshot 2025-11-10 at 1.04.35β€―PM.png) > [!note]- πŸ—‚οΈ Using Trained Documents? > Enables the mentorAI to use materials connected in the **Datasets** section. > **Use this for:** Allowing the mentor to reference specific documents or files (e.g., campus guides, program FAQs, or academic resources). > **Tip:** Keep this enabled if your mentor is meant to give context-specific guidance based on uploaded or approved materials. > [!note]- 🌐 Using Web Search? > Allows the mentorAI to search the internet for up-to-date information beyond its trained data. > **Use this for:** Providing current information, resources, or examples that may not be in the dataset. > **Tip:** Turn this off if you want the mentor to rely only on approved Syracuse-related materials. --- ## **🧩 Memory** The **Memory tab** is designed to let mentorAI store and reference helpful information from past interactions. ![](./attachments/567279621/Screenshot 2025-11-10 at 1.04.44β€―PM.png) > [!note]- πŸ—‚οΈ Memory Categories > Each saved memory can be sorted into a category for better organization: > > - **Personal Information:** Notes about user context, preferences, or goals (only when appropriate and privacy-safe). > - **Lessons Learned:** Reflections or takeaways from past interactions. > - **Help Requests:** Information about recurring questions or support topics students ask for. > - **Knowledge Gaps:** Areas where users often need clarification, helping mentors improve future responses. > [!note] > #### This feature is currently experimental. As mentorAI continues to evolve, memory will allow mentors to adapt to student needs more effectively while maintaining Syracuse Universitys privacy and data protection standards. --- ## **πŸ•“ History** The **History tab** allows you to view and manage all past conversations with your mentorAI. This includes reviewing chat logs, filtering interactions by topic or sentiment, and tracking overall engagement trends. ![](./attachments/567279621/Screenshot 2025-11-10 at 1.04.54β€―PM.png) > [!note]- πŸ“Š Key Features > 1. **User Search:** Filter past conversations by specific users. > 2. **Date Range:** Narrow results to a specific period for analysis. > 3. **Sentiment Topics:** Review how students felt about the conversation and what subjects came up most often. > 4. **Export Option:** Download interaction data for review or reporting. > 5. **Rating System:** Displays an overall score (out of 5). A higher rating indicates better user satisfaction and response quality. **Note:** New mentorAIs will have no conversation data or ratings are available. Once more users begin using the mentor, this section will increasingly help track engagement and mentor effectiveness --- ## **πŸ“Š Datasets** The **Datasets tab** allows you to manage the training materials and knowledge sources your mentorAI can reference while responding to users. By connecting files, links, or cloud resources, you can help the mentor provide accurate, context-specific guidance for students. ![mentorai datasets panel showing list of urls and resources with toggle switches turned on](./attachments/567279621/Screenshot 2025-11-10 at 1.05.14β€―PM.png) ### πŸ“„ Supported Data Types | **Type** | **Function** | **Limitations / Notes** | | --- | --- | --- | | **PowerPoint / PDF / DOCX / TXT / Excel** | Pulls readable text, titles, and notes to use as knowledge references. | Cannot read images, charts, or handwritten notes inside slides. | | **Google Drive / OneDrive / Dropbox** | Connects directly to stored files in the cloud for continuous access. | Access permissions must allow mentorAI to view the file. | | **URL / Web Crawler** | Uses website content to gather updated information or academic resources. | Must be a text-based site (not image-heavy or behind a login). | | **YouTube / Video / Audio** | Uses transcripts or captions to extract educational or research-related information. | Requires transcript or subtitles to tokenize. | | **GitHub** | Useful for research or coding-related projects where the mentor references repositories. | Reads text-based files only (like `.py`, `.md`, or `.txt`). | | **Image / ZIP** | Reads any text metadata or OCR-extracted text (if available). | Cannot analyze the visual content itself. | | **Web Crawlers** | Scans linked pages within a provided URL to gather additional related content. | Only captures textual updates, not images or media. | > [!info] > **πŸ’‘ Tip:** To make the most of your dataset uploads, prioritize resources with **text-based content** like websites, PDFs, and guides. This ensures the mentors responses stay relevant and academically grounded. --- ## **πŸ–₯️ Embed** The **Embed** section allows you to configure how your mentorAI appears and functions when placed on a website or within another digital platform. This includes visual customization, permissions, and interactive features. ![mentorai embed panel showing options for context aware single sign on and shareable link](./attachments/567279621/Screenshot 2025-11-10 at 1.12.34β€―PM.png) ### βš™οΈ Core Features | **Setting** | **Function** | **Notes / Example** | | --- | --- | --- | | **Advanced CSS** | Lets you style the mentor widget with custom CSS (colors, fonts, borders, etc.). | For advanced customization ideal for matching SUs brand style. | | **Icon Selection** | Choose between a default icon or upload a **custom icon** (logo or mascot). | The icon shown in the **Live Preview** at the bottom right corner. | | **Mode Selection** | Switch between **Default** and **Advanced** modes. | Default = simple chat only. Advanced = adds tabs like *Summarize*, *Translate*, and *Expand*. | | **Who Can View** | Restrict visibility to **Students**, **Administrators**, or **Anyone**. | Most mentors should be set to **Students** (all university members are considered students in MentorAI). Select this option and the Who Can Chat option to **Anyone** to allow anonymous users, useful for public or website chatbots. *(Options may change in future updates.)* | | **Who Can Chat** | Limit who can actively interact: **Anyone** or **Authenticated Users** (those signed in). | Useful if your site requires SU NetID or another login. | | **Website URL** | The web address where the mentor will be embedded. | Used for generating a unique embed token tied to that domain. | | **Get Token** | Generates an authentication token that allows the chat to run securely on that URL. | Likely required for linking the embed to an official SU page. | | **Context Aware** | Enables the mentor to use the website context (like the page topic) when answering. | Good for mentors hosted on resource pages or help sites. | | **Single Sign-On (SSO)** | Integrates with campus authentication so users dont need to log in separately. | May be used by the IT department. | | **Open by Default** | Makes the chat window automatically open when a user loads the page. | Helpful for welcome messages or proactive engagement. | | **Show Attachment** | Adds an option to upload or attach files directly in chat. | Ideal for sending PDFs, docs, or screenshots. | | **Show Voice Call / Record** | Enables audio interaction either real-time calling or voice message recording. | Still experimental in many mentor setups. | | **Shareable Link** | Generates a public link to share the chat outside the embedded site. | Useful for testing or sharing with teammates. | | **Create Embed** | Finalizes all chosen settings and generates a **code snippet** (HTML `script` or `iframe`) for embedding on your chosen website. | This is code that a web developer or IT admin can use to add the mentor to a website. | --- ## ❓ Frequently Asked Questions (FAQ) > [!note]- Can MentorAI access or remember personal information from users? > No. mentorAI does not automatically store personal data. Any saved memory must be manually added under the **Memory** tab by an administrator or creator/owner of a mentor. > [!note]- Why isnt my dataset working or showing up in responses? > mentorAI can only process text-based content. If your file or link mostly contains images or lacks visible text (like a scanned PDF), it wont be tokenized properly. > [!note]- What does token count mean in the Datasets section? > Tokens represent chunks of readable text that the AI has processed. More tokens generally mean more data for the mentor to reference. > [!note]- Whats the difference between Web Search and Trained Documents? > *Trained Documents* come from your uploaded datasets, while *Web Search* lets the mentor pull current information from the internet when responding. --- title: "mentorAI @ Syracuse University" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-syracuse-university.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-syracuse-university.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/544505857/mentorAI+@+Syracuse+University" page_id: "544505857" tags: ["mentorai", "overview", "teaching-learning", "study-tools", "getting-started"] audience: ["students", "faculty", "staff"] last_modified: "2025-12-11" --- # What is mentorAI The mentorAI platform is a Syracuse University run AI platform that enables AI-powered personalized assistants for students, faculty, and staff. [Try Now!](https://mentor.ai.syr.edu/) (mentor.ai.syr.edu) ## πŸŽ“ For Students: Personalized, Always-On Learning **Create Your Own Study Mentor -**Build AI-powered tutors tailored to your learning goals, courses, and textbooks. **Adaptive Learning Paths -**The AI adjusts to your pace and understanding, providing targeted feedback and curated practice questions. **Verified Knowledge, Not Guesswork -**Every answer is grounded in your course materials, with transparent citations and relevant sources. πŸ“ˆ **Benefits -**Higher engagement, deeper understanding, and better academic outcomes. ## πŸ‘©πŸ« For Faculty: Control, Efficiency, and Empowerment **Your AI Teaching Assistant -**Offload repetitive questions, grading prep, and FAQs to custom course-specific AI mentors. **Review Feedback -**Save time with AI-assisted assignment review and feedback, while maintaining full oversight. **Curriculum Design Support -**Generate draft lesson plans, quizzes, and materials in minutes. **Quality Control -**Each mentor uses faculty-approved content only, ensuring integrity and accuracy. 🧠 **Benefits -** More time for teaching and research, Ability to focus on the things that enable deeper understanding with your student. --- ## mentorAI Links [mentorAI Creating A Mentor](./mentorai-creating-a-mentor.md) [mentorAI Settings Options](./mentorai-settings-options.md) [mentorAI Setting Up a Web Crawler](https://answers.atlassian.syr.edu/wiki/search?text=mentorAI+Setting+Up+a+Web+Crawler&spaceKey=ITSAI) [mentorAI - Using the API](./mentorai-using-the-api.md) [mentorAI platform documentation](https://docs.ibl.ai/) --- title: "mentorAI - Tools" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-tools.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-tools.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/700121089/mentorAI+-+Tools" page_id: "700121089" tags: ["mentorai", "how-to", "teaching-learning"] audience: ["faculty", "staff"] last_modified: "2026-01-07" --- ## Overview mentorAI has many tools available for each mentor. These tools can add new functionality to a mentor and there are more that may be rolled out in the future. This page is meant to describe these tools --- ## Enabling a Tool To enable a tool, follow these steps: 1. Log into [mentor.ai.syr.edu](http://mentor.ai.syr.edu) 2. Navigate to the mentor you have permission to edit 3. Under the drop down for the mentor, go to Tools ![screenshot showing the dropdown menu and location of tools highlighted with a red box](./attachments/700121089/image-20260107-144822.png) 4. In the Tools settings, enable the tool. Below is a screenshot with a red box showing where to enable a tool. ![screenshot ](./attachments/700121089/image-20260107-145352.png) 5. The tool should now be enabled. Some tools require additional steps, outlined below. ### Additional Steps for some Tools Some tools need to be turned on at the prompt level as well, these include: - Code Interpreter - Image Generation - Screen Share Once the tool is enabled in the Tools setting, navigate back to the mentor chat screen and enable the tool under the chat box. ![Screenshot showing the box that needs to be activate on a new chat window, highlighted by a red box](./attachments/700121089/image-20260107-150213.png) --- ## Additional Considerations Some tools work best, or only function, with some LLM models. Feel free to explore which model is best for each tool and your use case. --- ## Image Generation This tool can generate basic images when given a prompt. Below is an example output: ![screenshot of a mento prompt asking it to generate a yellow duck on a lake. The mentor responded with the generated image of a duck on a lake](./attachments/700121089/image-20260107-151401.png) --- ## Code Interpreter This tool is helpful to visualize abstract math or data solutions. Some use cases include: - Plotting integrals, derivatives, Riemann sums, and more. - Comparing functions (e.g., yβ€―=β€―sinβ€―x vs. yβ€―=β€―cosβ€―x) to study amplitude, period, and phase shifts. - Graph experimental data (e.g., projectile motion) and fit trend lines or polynomials to validate theoretical models. - Upload small datasets (CSV or pasted values) and request histograms, box plots, or scatterplots to analyze distributions and correlations. ![screenshot showing the response from the mentor when asked to plot the integral of x-squared from 1 to 100. A graph is shown](./attachments/700121089/image-20260107-150123.png) --- title: "mentorAI - Using the API" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-using-the-api.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/clementine-platform/mentorai-using-the-api.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/591101962/mentorAI+-+Using+the+API" page_id: "591101962" tags: ["mentorai", "how-to", "api", "integration", "access"] audience: ["faculty", "staff", "IT"] last_modified: "2026-03-05" --- > [!note] > This document is still under construction, thank you for your patience. ## Overview This document is designed to provide starting point with interacting with mentorAI via its API and programmatically. --- ## Interacting with the API There are many ways to interact with an API and below are two examples, and how to set their environments up. ### VS Code Visual Studios Code is the light version of Visual Studios. It provides a powerful environment to program from and is the recommended IDE to use. [Setting up VS Code](https://answers.atlassian.syr.edu/wiki/x/BQA8Iw) --- ## API Methods In short, an API method is the basic idea of what you want an API to do. If all you wish to do is get information in regards to something you would use GET as the method. If you wish to create something new or request the API to make something, you would likely use a POST method. --- ## Getting an API Key > [!info] > Despite an API key appearing as an option on a mentor, it is actually tied to your account, not a specific mentor. The API key has the same permissions your user account does. At this time, request access to the API by emailing [aihelp@syr.edu](mailto:aihelp@syr.edu) with your username (netid) and please provide a brief description of what you plan to use the API for so we know different use cases for the product. 1. Afterwards, log back into mentorAI and navigate to a mentor you created or have been granted access to and select the API action in the drop down. ![Screenshot of a mentor's drop down menu, including the API selection near the bottom.](./attachments/591101962/image-20251118-141734.png) 1. Press the Create New button 1. Provide a name that is concise but descriptive of the keys use-case 2. Pick an expiration date. Common intervals are 1 month, 3 months, 1 year. 3. Press Submit 2. Copy the string that is now provided to you. This string will **NOT** be shown to you again. This string is your API key/token. Do **NOT** share it with anyone. Store the string somewhere safe, like a password manager. > [!note] > When programming, do **NOT** place your API key directly in your code. Instead, place it somewhere else and pull it into the code. Such as storing it in a .txt file and reading the file in the code. > > If you use GitHub, be sure to not accidently upload your key into your repository! --- ## API Documentation IBL (the creators of mentorAI) have a documentation resource that you may find useful. See the links below: [ibl-data-manager (Public)](https://docs.ibl.ai/apis/ibl.3.59.0/) --- ## Examples Currently the examples are in the PowerShell language. > [!info] > In the future, our examples may be moved to GitHub or another method, and links will be provided. This will be to offer more programming language support for each example. ### Get Mentors To get the mentors you have created, you will need to interact with a separate endpoint at this time. It is commented out in the code block below and shown here: ``` $urlPathway = https://base.manager.ai.syr.edu/api/ai-search/personalized-mentors/?platform_key=$orgId ``` ```powershell $ErrorActionPreference = 'Stop' #=== GET API KEY FROM FILE === [System.String]$apiKey = (Get-Content -Path "C:\path\to\your\apikey.txt" -Raw).Trim() #=== BUILD REQUEST HEADERS === $headers = @{ "Content-Type" = "application/json" "Authorization" = "Api-Token $apiKey" } [System.String]$orgId = "syracuse" [System.String]$userId = "{INSERT YOUR NETID HERE}" #can also use your FID #=== API URL TO GET ALL MENTORS FOR A USER === # Below are multiple methods and examples of API endpoints to accomplish this goal # Get the list of mentors you have access to. #$urlPathway = "https://base.manager.ai.syr.edu/api/ai-mentor/orgs/$orgid/users/$userid/" # Get the list of mentors you have access to, but pre-sort the response by name of the mentor #$urlPathway = "https://base.manager.ai.syr.edu/api/ai-mentor/orgs/$orgid/users/$userid/?filter_by=name" # Get the list of mentors created by a user (may require a user's FID as opposed to their netid) #$urlPathway = "https://base.manager.ai.syr.edu/api/ai-search/mentors/?platform_key=$orgId&created_by=$userid" # Similar to the first example, Get a list of mentors you have access to $urlPathway = "https://base.manager.ai.syr.edu/api/ai-search/mentors/?platform_key=$orgId" # Get the specific mentors you have created #$urlPathway = "https://base.manager.ai.syr.edu/api/ai-search/personalized-mentors/?platform_key=$orgId" #=== CALL THE API === try { $response = Invoke-RestMethod -Uri $urlPathway -Method Get -Headers $headers $response.results | Format-List } catch { Write-Host "StatusCode:" $_.Exception.Response.StatusCode.value__ Write-Host "StatusDescription:" $_.Exception.Response.StatusDescription Write-Host $_.Exception exit } ``` ### Get Mentor Datasets ```powershell $ErrorActionPreference = 'Stop' #=== GET API KEY FROM FILE === [System.String]$apiKey = (Get-Content -Path "C:\path\to\your\apikey.txt" -Raw).Trim() #=== BUILD REQUEST HEADERS === $headers = @{ "Content-Type" = "application/json" "Authorization" = "Api-Token $apiKey" } [System.String]$orgId = "syracuse" [System.String]$userId = "{INSERT YOUR NETID HERE}" #can also use your FID [System.String]$pathway = "{INSERT MENTOR STRING HERE}" #You can find this at the end of the URL string when viewing a mentor in a browser. Will be formatted like: 12345678-1234-1234-1234-123456789012 #=== API URL TO GET ALL DATASETS FOR A PATHWAY === $urlPathway = "https://base.manager.ai.syr.edu/api/ai-index/orgs/$orgid/users/$userid/documents/pathways/$pathway/" #=== CALL THE API === try { $response = Invoke-RestMethod -Uri $urlPathway -Method Get -Headers $headers $response.results | Format-List } catch { Write-Host "StatusCode:" $_.Exception.Response.StatusCode.value__ Write-Host "StatusDescription:" $_.Exception.Response.StatusDescription Write-Host $_.Exception exit } ``` ### Upload a document to a Mentor (Upload a dataset) ```powershell $ErrorActionPreference = 'Stop' #=== GET API KEY FROM FILE === [System.String]$apiKey = (Get-Content -Path "C:\path\to\your\apikey.txt" -Raw).Trim() #=== BUILD REQUEST HEADERS === $headers = @{ "Content-Type" = "application/json" "Authorization" = "Api-Token $apiKey" } [System.String]$orgId = "syracuse" [System.String]$userId = "{INSERT YOUR NETID HERE}" #can also use your FID [System.String]$url = "https://base.manager.ai.syr.edu/api/ai-index/orgs/$orgId/users/$userId/documents/train/" [System.String]$pathway = "{INSERT MENTOR STRING HERE}" #You can find this at the end of the URL string when viewing a mentor in a browser. Will be formatted like: 12345678-1234-1234-1234-123456789012 [System.String]$ingestURL = "{ENTER URL OF PUBLIC WEBSITE YOU WISH TO UPLOAD HERE}" $body = @{ "pathway" = $pathway "url" = $ingestURL "type" = "url" "access" = "private" } | ConvertTo-Json #=== CALL THE API === try { $response = Invoke-RestMethod -Uri $url -Method Post -Headers $headers -Body $body } catch { Write-Host "StatusCode:" $_.Exception.Response.StatusCode.value__ Write-Host "StatusDescription:" $_.Exception.Response.StatusDescription Write-Host $_.Exception exit } Write-Host $response.message Write-Host "Task ID:" + $response.task_id Write-Host "Doc ID: " + $response.document_id ``` ### Create chat session with Mentor > [!info] > Use this code below to get a sessionId. Then use the sessionId in the code to send a chat message. You can also get a sessionId from the developer tools in a browser. Steps are explained after the code for sending a message. ```powershell $ErrorActionPreference = 'Stop' #=== GET API KEY FROM FILE === [System.String]$apiKey = (Get-Content -Path "C:\path\to\your\apikey.txt" -Raw).Trim() #=== BUILD REQUEST HEADERS === $headers = @{ "Content-Type" = "application/json" "Authorization" = "Api-Token $apiKey" } [System.String]$orgId = "syracuse" [System.String]$userId = "{INSERT YOUR NETID HERE}" #can also use your FID [System.String]$mentorName = "{INSERT NAME OF THE MENTOR HERE}" [System.String]$url = "https://base.manager.ai.syr.edu/api/ai-mentor/orgs/$orgId/users/$userId/sessions/" $body = @{ "mentor" = $mentorName } | ConvertTo-Json #=== CALL THE API === try { $response = Invoke-RestMethod -Uri $url -Method Post -Headers $headers -Body $body $sessionId = $response[0].session_id } catch { Write-Host "StatusCode:" $_.Exception.Response.StatusCode.value__ Write-Host "StatusDescription:" $_.Exception.Response.StatusDescription Write-Host $_.Exception exit } ``` ### Send a chat message to a mentor > [!info] > We are working on trying to set it up so you can use the API to talk with a mentor rather than going through a web socket. Thank you for your patience > > We have an SSE connection as a possibility as well, this would still require an access code, but may be useful depending on your scenario. Please reach out to us if you are interested in an example to connect via SSE. ```powershell $ErrorActionPreference = 'Stop' [System.String]$sessionId = "12345678-1234-1234-1234-1234567890" # From the previous code [System.String]$orgId = "syracuse" [System.String]$userId = "{INSERT YOUR FID HERE}" # See information below this code snippet on how to obtain this [System.String]$pathway = "{INSERT MENTOR STRING HERE}" #You can find this at the end of the URL string when viewing a mentor in a browser. Will be formatted like: 12345678-1234-1234-1234-123456789012 [System.String]$prompt = "Hello, what's your name?" # Step 2: Create the JSON payload $payload = @{ flow = @{ name = $pathway tenant = $orgId username = $userId pathway = $pathway } session_id = $sessionId token = "{INSERT TOKEN STRING HERE}" # See information below this code snippet on how to obtain this prompt = $prompt } | ConvertTo-Json # Step 3: Create and connect a webSocket [System.Net.WebSockets.ClientWebSocket]$ws = New-Object System.Net.WebSockets.ClientWebSocket $uri = [System.Uri]::new("wss://asgi.data.ai.syr.edu/ws/langflow/") [System.Threading.CancellationTokenSource]$cts = New-Object System.Threading.CancellationTokenSource try { # Connect to WebSocket #Write-Host "Connecting to WebSocket..." -ForegroundColor Yellow $connectTask = $ws.ConnectAsync($uri, $cts.Token) $connectTask.Wait() #Write-Host "Connected!" -ForegroundColor Green # Step 4: Send the message #Write-Host "Sending message..." -ForegroundColor Yellow $bytes = [System.Text.Encoding]::UTF8.GetBytes($payload) $buffer = New-Object System.ArraySegment[byte] -ArgumentList @(,$bytes) $sendTask = $ws.SendAsync($buffer, [System.Net.WebSockets.WebSocketMessageType]::Text, $true, $cts.Token) $sendTask.Wait() #Write-Host "Message sent!" -ForegroundColor Green # Step 5: Receive responses #Write-Host "`nReceiving responses:" -ForegroundColor Yellow $fullResponse = "" $receiveBuffer = New-Object byte[] 4096 $segment = New-Object System.ArraySegment[byte] -ArgumentList @(,$receiveBuffer) do { $receiveTask = $ws.ReceiveAsync($segment, $cts.Token) $receiveTask.Wait() $result = $receiveTask.Result $message = [System.Text.Encoding]::UTF8.GetString($receiveBuffer, 0, $result.Count) $jsonResponse = $message | ConvertFrom-Json # Display the response #Write-Host $message # If it's a data chunk, accumulate it if ($jsonResponse.type -eq $null -and $jsonResponse.data) { $fullResponse += $jsonResponse.data } # Check if we've reached the end of responses if ($jsonResponse.eos -eq $true) { break } } while ($ws.State -eq [System.Net.WebSockets.WebSocketState]::Open) # Display the completed response #Write-Host "`n`nComplete Response:" -ForegroundColor Cyan Write-Host $prompt Write-Host $fullResponse } catch { Write-Host "Error: $_" -ForegroundColor Red } finally { # Step 6: Clean up if ($ws.State -eq [System.Net.WebSockets.WebSocketState]::Open) { $ws.CloseAsync([System.Net.WebSockets.WebSocketCloseStatus]::NormalClosure, "Done", $cts.Token).Wait() } $ws.Dispose() $cts.Dispose() #Write-Host "`nWebSocket closed." -ForegroundColor Yellow } ``` #### Getting a token and your FID We are working on a way to programmatically get a token, however, at this time, you can get one from your web browser. Follow the steps outlined below: > [!note]- Steps > 1. Navigate to [mentorAI](https://mentor.ai.syr.edu/platform/syracuse) > 2. Go to a mentor you are the creator or have permission to > 3. Open the developer tools for the web browser you are using > > 1. Such as `F12` or `Ctrl + Shift + I` on a Windows computer > 2. `Fn + F12` or `Command + Option + I` on a Mac computer > 4. Send a chat message to the mentor, you can ask it anything. > 5. Go to the Network tab filter on Socket (might be called WS, WebSocket, etc) select langflow/ > > 1. Make sure you are on the Messages tab to the right of langflow/ > 6. Look at a message and grab the FID (username) and the Token that was used. > > ![Screenshot of dev tools in a chrome browsers showing the buttons to press as outlined above. Specific information is redacted](./attachments/591101962/image-20251124-142630.png) --- title: "Copilot \u2013 Frequently Asked Questions" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/copilot/copilot-frequently-asked-questions.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/copilot/copilot-frequently-asked-questions.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/522289260/Copilot+\u2013+Frequently+Asked+Questions" page_id: "522289260" tags: ["copilot", "faq", "data-privacy", "m365", "security"] audience: ["students", "faculty", "staff", "IT"] last_modified: "2025-10-30" --- ### **Who owns the information I create or enter when using Copilot?** The user retains all ownership of any inputs to Microsoft Copilot and is granted ownership over any outputs that Copilot creates. --- ### **What types of university information are appropriate to use with Copilot?** When logged in with your SU credentials, you can use Copilot with all [classifications](https://answers.atlassian.syr.edu/wiki/search?text=Data+Classification+Definitions&spaceKey=infosec011) of university data that align with Syracuses data handling policies. --- ### **How long does Copilot retain my data or chat history?** Copilot conversations are ultimately stored in a users M365 mailbox in a hidden folder. The data will remain recoverable searchable by administrators for up to 30 days after a user deletes the conversation in Copilot, unless the mailbox has different retention policies applied. --- ### **Does Syracuse University have access to my Copilot activity or data?** Like all university technology systems, data within Syracuse Universitys Microsoft 365 (M365) environment may be accessed by authorized IT staff when necessary to support, maintain, secure, troubleshoot, or back up those systems. At Syracuse, this access is strictly governed by the [Information Technology Resources Acceptable Use policy](https://policies.syr.edu/policies/free-speech/information-technology-resources-acceptable-use-policy/) and the Information Security Framework. It is exercised in accordance with departmental guidelines and best practices. --- ### **Does Copilot train on my data?** Microsoft does **not** use your content, prompts, or Copilot interactions from Syracuse University accounts to train its AI models. All activity within SUs Microsoft 365 environment stays within the universitys secure tenant and follows Microsofts enterprise privacy commitments. Copilot offers[personalization and memory settings](https://learn.microsoft.com/en-us/graph/control-enhanced-personalization-privacy)that enable users and administrators to control how contextual data and prior interactions are utilized to enhance the experience. --- ### **Can Copilot be used in a private or temporary mode?** Copilot includes a feature called **Temporary Chat**, which lets you start a new conversation that isnt saved to your chat history. Once the window is closed or refreshed, the entire conversation is no longer visible, but the conversation is still bound by any retention policies that are applied to the user in M365. --- ### What do the Copilot buttons in Microsoft apps actually do? While the buttons are visible across many apps, their function doesnt extend beyond each programs workspace. They dont connect to or access additional work data outside your current file, document, or chat. Essentially, these buttons are shortcuts to the same Copilot experience built into each app rather than separate tools. --- title: "Google Gemini at Syracuse University" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/google-gemini-at-syracuse-university.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/google-gemini-at-syracuse-university.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/544538648/Google+Gemini+at+Syracuse+University" page_id: "544538648" tags: ["gemini", "overview", "getting-started", "productivity"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-30" --- # What is Gemini? Gemini is one of Syracuse University's AI assistants that helps you work smarter, not harder. Think of Gemini as your intelligent digital colleague who can help you write, analyze, research, and solve problems across virtually any subject area. ## Access Navigate to [Gemini](https://gemini.google.com/app) and login with your Syracuse University Google account (*netid*@g.syr.edu) ## Getting Started: The Basics The easiest way to get started is to ask Gemini a question or give Gemini a task. Here are a few simple ideas: - **Writing Assistant:** Help draft emails, reports, essays, and creative content - **Research Partner:** Summarize articles, explain complex topics, and find connections between ideas - **Problem Solver:** Break down complicated problems, suggest solutions, and help with analysis - **Learning Companion:** Explain concepts in simple terms, create study guides, and answer questions - **Data Analyst:** Help interpret data, create summaries, and identify patterns - **Creative Collaborator:** Brainstorm ideas, develop proposals, and refine your work ## How Does It Work? Using Gemini is as simple as having a conversation. You type your question or request in plain English (or other languages!), and Gemini responds with helpful, detailed answers. No coding or technical knowledge required. **Example interactions:** - "Help me write a professional email declining a meeting invitation" - "Explain quantum computing like I'm a freshman" - "What are the key points in this 50-page PDF I just uploaded?" - "Help me prepare for my presentation on climate change" - "Create a study schedule for my finals week" ## Best Practices for New Users **Start Simple:** Begin with straightforward questions to get comfortable with how Gemini responds. **Be Specific:** The more detail you provide, the better Gemini can help. Instead of "help with my essay," try "help me create an outline for a 5-page essay about the impact of social media on democracy." **Iterate and Refine:** Don't hesitate to ask follow-up questions or request changes. You can say things like: - "Make this more formal" - "Explain that in simpler terms" - "Give me three more examples" - "Make this shorter" **Verify Important Information:** While Gemini is highly capable, always double-check critical facts, citations, and data for important work. ## Showcase of Syracuse University Gemini uses --- --- title: "Google Gemini - Frequently Asked Questions" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/google-gemini-frequently-asked-questions.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/google-gemini-frequently-asked-questions.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/498597967/Google+Gemini+-+Frequently+Asked+Questions" page_id: "498597967" tags: ["gemini", "faq", "data-privacy", "getting-started"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-30" --- ### How do I switch from my personal Google account to my Syracuse University account to use Gemini? Sign in with your Syracuse University Google account, not your personal Gmail. When prompted, use your SU Google account, which is your **NetID (e.g., NETID@g.syr.edu)**. If youre already logged in with a personal account, click your profile icon **Add another account** sign in with your SU credentials. --- ### Will my chat data, prompts, and file uploads be used to train Google's AI models? User prompts, chats, and uploaded content used within Gemini for Education are not used to train Google's generative AI models. This is a core commitment of the enterprise-grade data protection under the [Google Workspace for Education Terms of Service Google Workspace](http://workspace.google.com/terms/education_terms/?utm_source=chatgpt.com). --- ### Is my chat data human-reviewed or accessed by Google personnel? Chats with Gemini in a Google Workspace for Education account are not human-reviewed or accessed by human reviewers to improve AI models. The data is managed with the same robust security controls as other core Workspace services (like Gmail and Drive). --- ### Does Google own the data I generate or upload to Gemini? You retain full control and ownership of your data. Your data is considered customer data, and Google only processes it according to the organization's instructions and terms. More information can be found on [Artificial Intelligence Guidelines](https://its.syr.edu/artificialintelligence/aiguidelines/). --- ### Does Syracuse University (SU) have access to the data, prompts, and information I input into AI platforms? At Syracuse, this access is strictly governed by the Information Technology Resources Acceptable Use policy ( ) and the Information Security Framework and is exercised in accordance with departmental guidelines and best practices. --- ### Can Gemini access other files in my Google Drive automatically? Gemini does **not** have unrestricted access to your entire Google Drive. It can only view or interact with files you explicitly open, reference, or share within a session. Unless you grant permission or directly include a file in your prompt, Gemini cannot access your other documents, folders, or emails. --- title: "How to use Google NotebookLM" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/how-to-use-google-notebooklm.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/how-to-use-google-notebooklm.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/530546723/How+to+use+Google+NotebookLM" page_id: "530546723" tags: ["notebooklm", "how-to", "study-tools", "teaching-learning", "data-privacy"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-24" --- NotebookLM helps you turn your course materials into insight. Upload your class readings, ask smart questions, and create summaries or study notes instantly. ## ⭐Setting Up Your First Notebooks 1. Click **New Notebook.** 2. **Add sources**: upload a Google Doc, PDF, or paste a website or YouTube link. 3. NotebookLM will analyze those files and display a short Source Guide with key topics and summaries. 4. Use the **Ask a Question** box to start chatting try asking things like: 1. Summarize this reading in 3 bullet points. 2. Compare the main arguments from each source. 3. What are the key takeaways from Chapter 3? > [!info] > ### What You Can Do > > - πŸ“š **Summarize Readings** Create short study notes or overviews of articles and case studies. > - 🧩 **Ask Targeted Questions** The model only uses the materials you upload (not the entire web). > - πŸͺ„ **Generate Learning Tools** In the *Studio* tab, turn your notes into flashcards, study guides, or even an audio recap. > - 🀝 **Collaborate** Share notebooks with teammates or classmates when working on a group project. --- ## Key Features and Capabilities of NotebookLM > [!note]- 🧭 Discover Sources > Find new information fast. Use the **Discover Sources** option to explore trusted web content related to any topic. You can even click **Im feeling curious** to have NotebookLM automatically build a random notebook. > > ![bcb269b8-8af6-480c-88d3-df5498dafade.png](./attachments/530546723/bcb269b8-8af6-480c-88d3-df5498dafade.png) > [!note]- βš™οΈ Configure Chat > Make the AI fit your study style. Under **Chat Settings**, choose between *Guided Learning*, *Default*, or *Custom* response modes and set replies to be **shorter or longer** depending on how detailed you want explanations to be. > **πŸ’‘ Pro Tip:** Using *Guided Learning* is ideal for studying or understanding new material. Use Customize to personalize how NotebookLM responds (e.g, Explain this like Im new to the topic) > > ![](./attachments/530546723/Screenshot 2025-10-24 at 2.34.03β€―PM.png) > [!note]- πŸ“ Save Notes > Save your favorite chat responses or add your own written notes right inside a notebook. You can even **convert your saved notes into a new source**, so NotebookLM treats your own writing as material it can summarize or connect to other readings. > πŸ’‘ **Pro Tip:** Use the πŸ‘ or πŸ‘Ž icons under NotebookLMs responses to give quick feedback. This helps the system refine future answers and better match your study style. > > ![](./attachments/530546723/Screenshot 2025-10-24 at 1.23.39β€―PM.png) > [!note]- πŸ”Š Audio Overviews > NotebookLM can automatically generate an **audio overview** of your notebooks content. Perfect for quick listening before class or while walking between buildings. > **πŸ’‘ Pro Tip:** Go into Interactive mode and ask A.I podcaster questions through your microphone. > > ![](./attachments/530546723/Screenshot 2025-10-24 at 2.40.12β€―PM.png) --- ## NotebookLM Studio Features The **Studio** in NotebookLM is where your notes and sources come to life. It offers a suite of creative tools to help you study, present, and understand your material in new ways. > [!note]- 🎧 Audio Overview > Turn your sources into an AI-generated podcast. Choose between formats like **Deep Dive**, **Brief**, or **Debate** to hear your material explained conversationally. > **Use Case:** Create a quick audio summary to review readings while walking to class or before an exam. > [!note]- πŸŽ₯ Video Overview > Generate a short video summarizing your notebooks key ideas. Each style highlights main themes or visuals from your sources. > **Use Case:** Use this to make class presentations more engaging or share visual overviews with group members. > [!note]- 🧠 Mind Map > Visualize how concepts, themes, or case findings connect across your sources. > **Use Case:** Perfect for brainstorming essays, identifying relationships between leadership theories, or mapping business strategy frameworks. > [!note]- πŸ“Š Reports > Compile your notebooks main ideas into a structured written summary or briefing document. > **Use Case:** Ideal for summarizing multiple readings or generating a one-page report for a research project or consulting case. > [!note]- πŸ—‚οΈ Flashcards > Automatically turn notes or responses into flashcards for review. > **Use Case:** Great for memorizing definitions, frameworks, or key takeaways before tests or presentations. > [!note]- ❓ Quiz > Let NotebookLM test your understanding with custom quiz questions based on your notebook content. > **Use Case:** Use this to check comprehension after uploading class readings or before a group discussion. > [!info] > ### πŸ’‘ Pro Tip > > The Studio tools work best after youve uploaded multiple sources and added notes or summaries. The more content NotebookLM can draw from, the more useful your audio, video, and quiz results will be. --- ## FAQ Limitations > [!note]- What source types can I upload to NotebookLM? > Source types supported by NotebookLM are: > > - **Google Workspace Files:** You can import **Google Docs** and **Google Slides**. > - **Document Formats:** Supported document file types include **Text files**, **Markdown documents**, and **PDF files**. NotebookLM has limited support for **image-only PDFs**. > > - **Web Content:** You can use **Web URLs** and **Copy and pasted text**. > > - When importing a Web URL, **only the visible text content** of the HTML webpage is scraped for use; images, embedded videos, or nested webpages are **not imported**. > - **Paywalled webpages** are not supported. > - **Multimedia:** NotebookLM supports **Audio files** and **YouTube URLs of public videos**. > > - For YouTube, only **public videos with captions** (user-uploaded or auto-generated) are supported, and **only the text transcript** is imported as the source. Videos without speech are not supported. > - Local audio files (such as **MP3** and **WAV**, among others) are **transcribed at the time of import**, and the resulting text is saved as the source. Audio imports may fail if the audio is of low quality or does not contain speech. > [!note]- How many sources can I upload, and what are the file size limits? > NotebookLM places strict technical limitations on both the overall number of sources per notebook and the size of each individual source. > > Limits for Standard Users > > 1. **Source Count Limit:** You can include up to **50 sources per notebook**. > > 1. **File Size Limits per Source:** Each individual source has two primary capacity restrictions: > > 1. **Word Count:** Up to **500,000 words** per source. > 2. **File Size (for uploads):** Up to **200MB** for local uploaded files. > [!note]- Can I share my notes or notebooks with other students? > Yes, NotebookLM supports sharing and collaboration, although there are restrictions based on your account type. > > - **Sharing Study Aids:** You can create study sets like **flashcards and quizzes** from your documents and share these with friends or colleagues using a **simple link**. > > - **Notebook Sharing Restrictions:** The notebooks can **only be shared within Syracuse University's domain.** > > - **Real-time Collaboration:** When a notebook is shared, collaborators who have editor permissions can work together, and edits to notes within that shared notebook appear **in real-time**. > > - **Advanced Features:** Advanced sharing features may be available with premium editions. > [!note]- How is my personal data and uploaded content handled for privacy? > NotebookLM is covered by the [Google Workspace for Education Terms of Service](https://workspace.google.com/terms/education_terms/). These terms include **enterprise-grade security and privacy controls**. > > Specifically, under these terms: > > - Your uploaded files, chats, and model outputs **will not be reviewed by human reviewers**. > > - They **will not be used to improve generative AI models**. > > - The service also supports compliance with industry regulations like [FERPA and COPPA](https://educationframework.com/resources/student-privacy-laws/federal-laws). **NotebookLM Help:** Visit [NotebookLM Help](https://support.google.com/notebooklm/?hl=en#topic=16164070) for tutorials, FAQs, and troubleshooting support. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu) --- title: "Smart Study Companion with Gemini" url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/smart-study-companion-with-gemini.html" markdown_url: "https://julianhernandez2155.github.io/su-kb-site/data-ai/gemini/smart-study-companion-with-gemini.md" source_url: "https://answers.atlassian.syr.edu/wiki/spaces/ITSAI/pages/515801118/Smart+Study+Companion+with+Gemini" page_id: "515801118" tags: ["gemini", "how-to", "study-tools", "teaching-learning"] audience: ["students", "faculty", "staff"] last_modified: "2025-10-30" --- Transform how you learn with a personalized Study Gem. Upload class materials and let Gemini explain difficult concepts, quiz you on topics, and create custom study guides all from your SU Google account. # Create a Personalized Tutor Gem **Youll need:**Course materials (PDFs, slides, lecture notes, or study guides) #### Steps: 1. From the left sidebar, select **Explore Gems** to browse or create your own AI helper. 2. Click **+ New Gem.** Give it a descriptive name like *Smart Study Companion* or *BIO 121 Tutor*. 3. Add instructions in the description and instruction fields to guide your Study Gem. - *Example:*You are my personal study tutor. Explain concepts in simple terms, create 5-question quizzes, and summarize uploaded lecture slides. 4. Upload your lecture notes, slides, or readings. Then ask: - Create a quiz for Chapter 2, or Explain the main themes from this document. --- πŸ’¬ *Example prompts:* > [!note]- Generate 5 multiple-choice questions from these notes. > Generate **5 multiple-choice questions** based on this document or lecture content. > > For each question: > > - Provide **4 answer options** (AD). > - Indicate the **correct answer** and include a **one-sentence explanation**. > - Mix conceptual, definition-based, and applied questions. > - Keep the questions at a college level (not overly simple). > > At the end, include a short summary titled **Concepts Reviewed** listing the key topics covered in the quiz. > [!note]- Summarize this article in 3 main points. > Summarize this article into **3 main points** that capture the key arguments or findings. > > For each main point: > > - Provide a **23 sentence explanation**. > - Highlight any significant data, evidence, or examples that support it. > > Conclude with a short **Overall Insight** section (2 sentences) that explains how these points connect or why they matter in context. > [!note]- Explain this formula step by step like a tutor. > Explain this formula **step by step** as if you were tutoring a student: > > 1. **Name and Purpose:** Briefly describe what the formula is used for. > 2. **Define Variables:** List and define each variable or symbol in plain language. > 3. **Step-by-Step Breakdown:** Explain how to apply the formula, showing an example calculation if possible. > 4. **Common Mistakes:** Identify typical errors students make when using this formula. > 5. **Check for Understanding:** End with one short practice question the student can solve. > > Keep explanations clear, structured, and easy to follow. > [!note]- Create a flashcard set for this topic. > Create a **flashcard set** based on this topic or document. > > Each card should have: > > - **Front:** A key term, concept, or question. > - **Back:** A concise definition or 12 sentence explanation. > > Generate at least **10 flashcards** that cover the main ideas, terms, and examples. > > Format them as: > Front: [Term/Question] > Back: [Definition/Answer] > > End the set with a short **Review Tip** suggesting how to use the cards for effective studying. ![A screenshot of the user interface for Gemini where it was provided a powerpoitn and asked to create multiple choice questions from it](./attachments/515801118/Screenshot 2025-10-17 at 4.31.47β€―PM.png) > [!info] > ### **Tips for using Tutor Gem** > > - Be specific tell Gemini what level of detail you need (college-level summary, 2-minute explanation, etc.). > - Reuse your Study Gem before every exam to keep your learning consistent. > - Edit or update your Gem anytime as you add new course content. > - Save and share a general Study Guide Gem link with classmates (see below). --- ### **Share a Sample Study Guide Gem** If youve built a helpful general study Gem, you can share it with others. Click **Explore Gems Click the Share Icon Copy Link**to generate a URL. Post the link in your **Blackboard**, **Google Classroom**, or **student group chat** for others to try. *Example:*[*SU General Study Guide Gem*](https://gemini.google.com/gem/1Ouhq6qtPEdbMvDxORaqDCJfIwEji7kKb?usp=sharing)*helps generate summaries, flashcards, and quizzes for any subject.* --- ## FAQ Limitations > [!note]- When I share a New Gem, do other users get access to my personal data or materials? > When you share a Gem link, others can view or use your *Gems setup* (its name, instructions, and structure), but they do **not** have access to your uploaded files, past chats, or custom content. Each users data remains private under their own SU Google account. > [!note]- How do I fill out the criteria for a new Gem? > Each Gem includes sections for **Name**, **Description**, **Instructions**, and **Knowledge** (uploaded files). For help writing effective instructions, refer to: > πŸ‘‰ > [!note]- What kinds of files can I upload to my Gem? > You can upload common study file types like **.pdf, .docx, .txt, .pptx, or Google Docs** directly from your Drive. Avoid uploading restricted or confidential data such as graded assignments, exams, or student records. ### **Keep in Mind** - Avoid uploading confidential or graded materials. - Review AI-generated quizzes for accuracy. - Always double-check facts Gemini is a learning support tool, not a replacement for understanding. --- ## Support Need help? **ITS Help Desk**: 315-443-2677 [help@syr.edu](mailto:help@syr.edu)