Questions
- What is LibreChat?
- Who can use LibreChat at Bowdoin?
- Is LibreChat secure? Can I put sensitive data into it?
- How do I choose an AI model in LibreChat?
- What is the context window and how does it affect my conversations?
- What is a prompt template and how is it different from custom instructions?
- Can I share or save my LibreChat conversations?
- How do I cite AI-generated content from LibreChat?
- What file types can I upload to LibreChat?
Environment
This article applies to all current Bowdoin College faculty, staff, and students with access to LibreChat. LibreChat is Bowdoin's AI chat platform, operated and managed by Information Technology and accessible at librechat.bowdoin.edu.
Resolution
LibreChat basics
What is LibreChat?
LibreChat is an AI chat platform operated by Bowdoin IT that provides access to multiple large language models (LLMs) — including OpenAI GPT models and Anthropic Claude — through a single Bowdoin login. Unlike public AI tools, LibreChat is managed by Bowdoin, meaning your conversations are protected under Bowdoin's data governance policies and your data is not used to train the underlying AI models.
What makes LibreChat different from public AI tools like ChatGPT?
LibreChat gives you access to many of the same underlying AI models as public tools, but within a Bowdoin-managed environment. Key differences include:
- Sign-in is through Bowdoin SSO — no separate account creation needed
- Your data is not used to train AI models
- LibreChat is approved for use with sensitive data under Bowdoin's data classification policy
- IT can monitor for policy compliance and support you if something goes wrong
Who has access to LibreChat?
All current Bowdoin faculty and staff have access. Students must first complete the Generative AI Ethics course in Canvas before access is activated. See Getting Started with LibreChat in the Related Articles section for login instructions.
Data security and privacy
What type of data is safe to put into LibreChat?
LibreChat is approved for use with sensitive data under Bowdoin's data classification policy. However, always apply good judgment: avoid sharing personally identifiable information about individuals who have not consented, and follow any departmental or research data handling requirements that apply to your work.
How secure is the data I put into LibreChat?
Data you enter into LibreChat is not used to train the underlying large language models. Conversations are stored within Bowdoin's managed environment and subject to standard IT data governance policies, including a 30-day retention period. For full details, see How Bowdoin Bot Works in the Related Articles section.
Models and features
What AI models are available in LibreChat?
LibreChat offers a range of models from multiple providers, including OpenAI GPT models, Anthropic Claude, Amazon Nova, Meta LLaMA, and DeepSeek. The default model is GPT-5. The available model list is updated as new models are added and older ones are retired. See Choosing an AI model in LibreChat in the Related Articles section for a current comparison.
What does selecting a model mean, and how do I pick one?
Different models have different strengths — some excel at creative writing, others at coding or structured reasoning. You select a model from the drop-down in the upper-left corner of the chat window. If you are unsure, start with the default (GPT-5) and experiment. See Choosing an AI model in LibreChat for guidance.
What does "temperature" mean in LibreChat?
Temperature controls how creative or unpredictable the AI's responses are. A lower temperature (closer to 0) produces more focused, consistent responses. A higher temperature (closer to 1 or above) produces more varied, creative responses. For most tasks, the default setting works well. See Adjusting model parameters in LibreChat for details on all parameters.
What file types can I upload to LibreChat?
LibreChat supports common file formats including PDFs, images, Word documents, spreadsheets, and text files. Uploaded files can be used as reference material for your conversation — the AI will draw on their contents when generating responses. See Using Files and Artifacts in LibreChat for full instructions.
Prompts, templates, and instructions
What is a Prompt Template?
A Prompt Template (also called a Prompt in LibreChat) is a reusable prompt you save to the Prompt Library. You can include placeholder variables (for example, {{topic}}) that are filled in each time you use the template. Templates are useful for tasks you run repeatedly with different inputs. See Manage prompts in LibreChat for instructions.
What are Custom Instructions?
Custom Instructions are persistent preferences you set for how the AI should respond to you across all conversations — for example, always responding formally, always assuming a specific level of expertise, or always including citations. They apply globally to your account rather than to a single conversation.
What is the difference between Prompt Templates and Custom Instructions?
- Prompt Templates — reusable prompts for specific tasks; applied to individual conversations when you choose to use them
- Custom Instructions — global preferences that apply to every conversation automatically
Context window and conversation limits
What is the context window?
The context window is the maximum amount of text an AI model can "see" at one time — including everything you have typed and everything it has responded with in the current conversation. Once a conversation exceeds the context window, earlier parts of the conversation may no longer influence the model's responses.
What is the limit on how much data I can input into a conversation?
Context window limits vary by model. As a rough guide, 1,000 tokens is approximately 750 words. Most current models support very large context windows (100,000 tokens or more), but very long conversations or large uploaded documents can approach these limits. You can adjust the Max context tokens parameter in the right-hand panel. See Adjusting model parameters in LibreChat for details.
Sharing and saving
Can I share and save conversations?
Yes. You can share a conversation by generating a link accessible to other Bowdoin SSO users, export individual conversations in multiple formats (.txt, .md, .json, .csv, .png), and save AI responses as notes within the chat. See Sharing and Exporting LibreChat conversations and Manage LibreChat chat history in the Related Articles section.
Citation guidance
How do I cite that I used a large language model?
Citation practices for AI-generated content are evolving. Most style guides recommend disclosing AI use and noting the model, date, and prompt used. Common examples:
- APA: OpenAI. (Year). ChatGPT [Large language model]. https://chat.openai.com — adapt as appropriate for the model used in LibreChat
- MLA: "Prompt text." Model Name, version, Publisher, Date, URL.
Always follow your instructor's or journal's specific guidance on AI citation. When in doubt, disclose AI use explicitly in your work.
Additional Help
If you need further assistance, you have several options:
- Bowdoin Bot: Chat with Bowdoin Bot directly from any KB page for instant answers.
- Phone: Call the Bowdoin College Service Desk at (207) 725-3030.
- In person: Visit the Tech Hub in Smith Union during business hours.
- Submit a ticket: Request assistance through the Service Catalog.