Here is the thing most people miss at the start: Claude begins every chat knowing nothing about you. Not your job, not your team, not who you are writing for, not the tone your company uses. It is like a brilliant new hire on day one, sharp as anything, but with zero idea how your world works.
Context is how you brief that new hire. It is the single biggest lever on quality after prompting itself, and it is easy. Ask a cold Claude to "write a follow-up email" and you get something generic. Tell it you are a logistics account manager, the client is annoyed about a late delivery, keep it warm but firm, three short paragraphs, and you get something you could almost send. Same request, night-and-day result, and the only thing that changed was the context.
Give it the material
The most direct context is the stuff you already have. Upload it. Claude reads PDFs, Word documents, text files, spreadsheets (CSV and Excel), and images, up to 30 MB a file, using the + button in the message box. Point it at the actual report, the real spreadsheet, the brief you were sent, and it works from the source instead of your description of the source.
This beats summarising from memory every time. If the answer depends on a document, hand Claude the document. Do not retype it, do not paraphrase it, do not hope Claude can guess. Give it the real thing and ask your question about it.
One tip on formats. Claude reads all of these happily, but they are not equal in weight. A PDF is read as a set of images, and a Word file carries a lot of hidden styling, so both spend far more of Claude's memory than the plain words would. Text and Markdown (a simple plain-text format, see the glossary) carry the same words with almost none of that overhead. So when you control the format and the material is large or you will reuse it often, prefer a .md or .txt file: same information, a fraction of the weight, and cleaner for Claude to read. When you do not control it, upload the PDF or Word file anyway. Any source beats no source.
Context isn't unlimited
Why does the weight matter? Because Claude has a context window: the amount it can hold in mind in a single chat, its working memory. It is large, but it is finite, not bottomless. Everything counts against it: the files you add, the conversation so far, and the answer Claude is writing.
Two things follow. First, long chats and heavy files fill it up. As a chat nears the limit, Claude quietly summarises the earlier parts to keep going, and in that summary some detail can slip. Second, lean beats bloated. A tidy Markdown brief and a focused chat leave more room for the real work than a stack of heavy PDFs and a conversation that has wandered for an hour.
So the habits from earlier pay off here: one topic per chat, start fresh when you switch tasks (lesson 2.2), and feed Claude clean, lean material. The full mechanics, tokens, why long chats drift, and when to start over, open Part 4. For now, treat the context window as precious space and try not to waste it.
Three places to put context
Context can live at three levels, from "just for now" to "for good". Picking the right one saves you from repeating yourself.
This message. Paste or upload context right where you need it. It lasts for the one chat. Perfect for a one-off ("here's the doc, summarise it").
This project. A project keeps files, instructions, and chats together for a piece of ongoing work, and its instructions apply to every chat inside it. Projects are open to everyone, free accounts included (up to five). This is the Part 4 topic; here, just know it exists for when a single chat is not enough.
Everything. Your profile instructions (in Settings, under "Instructions for Claude") apply to every chat, automatically. This is where your durable facts live: your role, how you like answers, terms your industry uses. Set it once and stop repeating yourself. The rule of thumb: the more often you'd retype something, the wider you should set it.
Write a short brief once
The highest-value ten minutes you will spend is writing a short brief about how you work, then dropping it into your profile instructions (or a project). Keep it tight, a handful of lines. Answer who, what, and how.
That is all it takes. Something like: "I'm a marketing lead at a mid-size logistics firm. I mostly write client updates and internal briefs. My readers are busy and non-technical. I like a warm, plain tone, short paragraphs, no jargon. Never invent figures; ask me if you're unsure." Paste that once and every answer starts closer to right.
One caution, straight from the safety lesson: context is powerful, but the rules still apply. On a personal plan, think before you paste anything sensitive or confidential. The personal-versus-business divide from lesson 1.4 has not gone away just because the context box is handy.
Where this goes next
Good context and good prompting are two halves of the same skill. Context tells Claude who it's helping; the prompt tells it what to do. Part 3 is all about the prompt. Later, in Part 4, that little brief grows up into projects and skills, reusable, shareable ways to give Claude context and behaviour on tap. For now, the habit is enough: never make Claude guess what it could simply be told.
Try it
Write your own brief, the five lines from the card above, about your real role. Keep it under a hundred words. Now open Settings and paste it into "Instructions for Claude".
Then test it. Start a fresh chat and ask for something you actually need ("draft a short update to my team about this week"). Notice how much closer the first answer lands now that Claude knows who you are. That gap, between a cold Claude and a briefed one, is the whole point of this lesson.
(General exercise. Role and industry versions come once accounts are in.)