Creating a Claude Project (Anthropic)

Claude is another advanced Large Language Model (LLM) (similar to ChatGPT and Gemini) developed by a company called Anthropic. You can access Claude at: https://claude.ai 

Similar to the other LLM tools covered in this course Claude is known for its ability to:

• Write and understand code effectively (great for automating classroom tasks or exploring educational tech)
• Handle repetitive tasks like rewriting text, translating instructions, or generating feedback
• Maintain context well during longer conversations

With Claude’ s Projects, you can save your instructions, documents, and ongoing work in one place. That means you don’t need to repeat your context every time you open a new chat.

Projects

Projects are organized workspaces on Claude.ai. Each Project has its own chat history, a knowledge base where you upload files (policies, handbooks, spreadsheets, code, etc.), and optional “project instructions” that shape Claude’s tone and behavior across every chat in that Project. This keeps context stable and reduces rework. On paid plans, you can also share Projects with colleagues (with “can use” or “can edit” permissions) so everyone works from the same materials. 

When you add lots of content, Projects automatically switch on retrieval-augmented generation (RAG). Instead of trying to stuff everything into the model’s window, Claude indexes your knowledge base and fetches only the most relevant passages at answer time. That means you can keep adding material and still get fast, high-quality responses, RAG can expand practical capacity by up to 10×.

Claude accepts common file types you likely use at work (PDF, DOCX, CSV, TXT, HTML, RTF, EPUB, JSON, and more) subject to size and token limits. For many PDFs under 100 pages, Claude can analyze both text and visuals (images, charts). With longer PDFs or other models, it will extract and reason over the text. 

A small but important detail: chats inside a Project don’t see each other unless you put the information into the Project’s knowledge base or instructions. That keeps conversations focused and avoids accidental cross-talk. 

Step-by-Step: Create a Project and Upload Files

1. Open Claude at claude.ai, hover over the left sidebar, and select Projects (or go to claude.ai/projects). Click + New Project.
2. Name your Project and add a short description for your own reference. If you’re on a Team or Enterprise plan, choose visibility (keep it private, share with specific members, or make it available to your organization).
3. Set Project instructions. Describe how you want Claude to behave: audience, tone, format, and any rules it should follow (for example, “write at a parent-friendly reading level,” or “summaries end with action items”). Save your instructions. Claude will apply them to every chat in this Project.
4. Add knowledge. In the Project’s right-hand knowledge panel, click + and upload your documents (policies, curriculum maps, templates, spreadsheets). Claude will process these and use them as context across the Project. If you add a lot, Claude will automatically enable RAG to scale capacity.
5. Start a chat inside the Project. Ask targeted questions like “Summarize Chapter 2 in five bullets with two discussion prompts,” or “Draft a parent letter based on the field-trip policy.” Reference file names to steer answers.
6. Iterate. If an answer feels generic, nudge Claude to cite the specific file or section, or refine your Project instructions (for example, “always quote policy language verbatim in a gray callout”).
   
7. (Optional) Share. On Team/Enterprise, click Share project to invite colleagues. 
   
8. Keep it tidy. Because only knowledge and instructions persist across chats, store key facts in the Project knowledge base rather than leaving them buried in one conversation. Archive or delete Projects when you’re done.