Claude Apps And Product Surfaces

The Core App

The Claude product overview presents Claude as an AI for problem solvers: a workspace for analysis, writing, coding, data work, and collaboration. The web, desktop, and mobile apps are the easiest place to use Claude interactively and to attach files or context without writing code.

The core app is where most product features appear first or become visible to non-developers: projects, artifacts, memory, connectors, voice or mobile access, and specialized workflows. Each feature has separate availability and privacy details, so verify the relevant support article before rollout.

Projects And Artifacts

Projects are durable workspaces with chat history, uploaded knowledge, and project instructions. They are useful when Claude needs stable context across related conversations: product specs, client research, legal background, recurring editorial voice, or codebase notes.

Artifacts are substantial outputs in a dedicated window: documents, code, visualizations, mini-apps, and other shareable content. They are useful when the deliverable needs iteration outside the chat transcript.

Connectors, Search, And Memory

Connectors let Claude access apps and services, retrieve data, and take actions within connected services. The support docs state that Claude inherits each person’s permissions from the connected service, which is the key governance point: connector access should mirror real organizational access.

Memory changed materially this quarter. Anthropic now describes one memory that works everywhere: what Claude remembers in chat is the same memory Cowork uses. You can view, edit, and delete memory at the topic level, and sensitive topics are off by default. Memory is on by default for Free, Pro, and Max; on Team and Enterprise it is admin-gated and off until the user enables it.

Incognito chats remain excluded from memory and chat search, subject to the retention caveats covered in the privacy guide.

Claude In Chrome

Claude in Chrome is now generally available on every paid Claude plan. It takes autonomous actions in the browser — navigating, filling forms, and completing multi-step tasks — with a safety classifier gating what it will do.

For organizations, the relevant control is the admin domain allowlist: Enterprise administrators can restrict which sites Claude is permitted to act on. Treat browser agency like connector access — scope it to the domains the work actually requires, and assume anything on an allowed page is readable context.

Specialized Claude Apps

Claude Cowork moves Claude from chat into delegated desktop work. Support docs describe Cowork as running directly on your computer with chosen files and tools, with code running in an isolated virtual machine but still capable of making real file changes. Cowork now shares the same memory as chat.

@Claude brings Claude into Slack threads for Team and Enterprise. Claude Science is the research workbench for scientific tooling and local or infrastructure-based compute, and it anchors the researcher seat program. Claude Design is the Anthropic Labs design surface. Claude for Teachers is the verified-educator track. Claude Security is a public beta for Enterprise security teams that scans code, validates findings, and suggests patches.

Tracking App Changes

App features ship faster than any guide can restate them. Claude Central deliberately does not run a weekly news feed — that lane belongs to claudeweekly.com, which covers Claude and Anthropic announcements week by week.

For dated feature changes, the Claude app release notes in the Help Center are the owner of that fact. Use this page for the shape of the product surface, the release notes for what shipped, and claudeweekly.com for context on why it matters.