The Claude app surfaces

Most of what people mean by “using Claude” happens in one chat box, and that’s fine — but the app has a few surfaces arranged around that box, and they change what the chat is good for. None of them require a terminal. None of them are separate things you buy. They’re modes and containers you switch on when the shape of the work calls for them.

The quickest way to keep them straight is by the complaint each one fixes. Artifacts fixes “I can’t tell what you actually made” — the result opens in a panel beside the conversation, where you can look at it and use it, rather than reconstructing it from a wall of text. Projects fixes “I keep re-explaining myself” — your files, your standards, and your standing instructions live in one workspace, and every chat started inside it inherits them. Deep Research fixes “that sounded confident, but where did it come from” — it spends minutes instead of seconds, runs many searches, reads the sources, and hands back a structured report with citations you can follow.

Projects is the one worth pausing on, because it is really context made permanent. Everything you would otherwise paste at the top of every conversation — who the audience is, what your house style forbids, which figures are current — becomes ambient instead of repeated. It’s the same instinct as a CLAUDE.md file in a code project, expressed as clicks rather than a file you edit.

There’s also a surface that isn’t inside the app at all: Claude in Slack, which you @mention in a channel the way you’d tag a colleague. The guide below covers what that changes for a team — mostly that the request and the result stay in the thread where the work is already being coordinated, instead of disappearing into one person’s private chat history.